# Auto 클래스

많은 경우, `from_pretrained()` 메서드에 전달된 사전 학습 모델의 이름이나 경로에서 사용할 아키텍처를 추론할 수 있습니다. [AutoConfig](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoConfig), [AutoModel](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel), 또는 [AutoTokenizer](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoTokenizer)를 인스턴스화하면 적절한 아키텍처 클래스가 자동으로 생성됩니다. 예를 들어:

```python
model = AutoModel.from_pretrained("google-bert/bert-base-cased", device_map="auto")
```

이것은 [BertModel](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertModel)의 모델 인스턴스를 생성합니다.

각 작업에는 해당하는 `AutoModel` 클래스가 있습니다.

## Auto 클래스 확장

각 auto 클래스는 커스텀 클래스로 확장할 수 있는 메서드를 제공합니다. 예를 들어, 커스텀 모델 클래스 `NewModel`을 정의했다면, 이에 해당하는 `NewModelConfig`도 함께 정의해야 합니다. 그런 다음 아래에 표시된 대로 auto 클래스에 등록할 수 있습니다:

```python
from transformers import AutoConfig, AutoModel

AutoConfig.register("new-model", NewModelConfig)
AutoModel.register(NewModelConfig, NewModel)
```

이후 auto 클래스를 평소와 같이 사용할 수 있습니다.

`NewModelConfig`가 [PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)의 하위 클래스인 경우, config를 등록할 때 사용하는 키(여기서는 `"new-model"`)와 동일한 값으로 `model_type` 속성이 설정되어 있는지 확인하세요.

마찬가지로, `NewModel`이 [PreTrainedModel](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel)의 하위 클래스인 경우, `config_class` 속성이 모델 등록 시 사용한 클래스(여기서는 `NewModelConfig`)와 동일하도록 설정되어 있는지 확인하세요.

## AutoConfig[[transformers.AutoConfig]]

#### transformers.AutoConfig[[transformers.AutoConfig]]

```python
transformers.AutoConfig()
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/configuration_auto.py#L286)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoConfig.from_pretrained) 클래스 메서드로 생성될 때 라이브러리의 구성 클래스 중 하나로 인스턴스화되는 범용 구성 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_pretrained[[transformers.AutoConfig.from_pretrained]]

```python
from_pretrained(pretrained_model_name_or_path: str | os.PathLike[str], **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/configuration_auto.py#L309)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다: - huggingface.co의 모델 저장소에 호스팅된 사전 학습된 모델 구성의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.save_pretrained) 메서드를 사용하여 저장된 구성 파일이 포함된 *디렉터리*의 경로, 또는 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained) 메서드, 예: `./my_model_directory/`. - 저장된 구성 JSON *파일*의 경로, 예: `./my_model_directory/configuration.json`.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 모델 가중치 및 구성 파일의 (재)다운로드를 강제하고, 캐시된 버전이 존재하는 경우 이를 덮어쓸지 여부.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

return_unused_kwargs (`bool`, *선택 사항*, 기본값 `False`) : `False`인 경우, 이 함수는 최종 구성 객체만 반환합니다. `True`인 경우, 이 함수는 `Tuple(config, unused_kwargs)`를 반환하며, 여기서 *unused_kwargs*는 키가 구성 속성이 아닌 키/값 쌍으로 구성된 사전입니다. 즉, `kwargs` 중 `config`를 업데이트하는 데 사용되지 않고 무시되는 부분입니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

kwargs(추가 키워드 인자, *선택*) : kwargs의 값 중 설정 속성에 해당하는 키는 로드된 값을 덮어쓰는 데 사용됩니다. 키가 설정 속성이 *아닌* 키/값 쌍에 대한 동작은 `return_unused_kwargs` 키워드 인자로 제어됩니다.

…

로드된 config 객체의 `model_type` 속성에 따라 인스턴스화할 configuration 클래스가 선택되며, 해당 속성이 없는 경우 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백됩니다:

- **EvollaModel** -- [EvollaConfig](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaConfig) (EvollaConfig 모델)
- **afmoe** -- [AfmoeConfig](/docs/transformers/v5.23.0/ko/model_doc/afmoe#transformers.AfmoeConfig) (AfmoeConfig 모델)
- **aimv2** -- [Aimv2Config](/docs/transformers/v5.23.0/ko/model_doc/aimv2#transformers.Aimv2Config) (Aimv2Config 모델)
- **aimv2_text_model** -- [Aimv2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/aimv2#transformers.Aimv2TextConfig) (Aimv2TextConfig 모델)
- **aimv2_vision_model** -- [Aimv2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/aimv2#transformers.Aimv2VisionConfig) (Aimv2VisionConfig 모델)
- **albert** -- [AlbertConfig](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertConfig) (AlbertConfig 모델)
- **align** -- [AlignConfig](/docs/transformers/v5.23.0/ko/model_doc/align#transformers.AlignConfig) (AlignConfig 모델)
- **align_text_model** -- [AlignTextConfig](/docs/transformers/v5.23.0/ko/model_doc/align#transformers.AlignTextConfig) (AlignTextConfig 모델)
- **align_vision_model** -- [AlignVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/align#transformers.AlignVisionConfig) (AlignVisionConfig 모델)
- **altclip** -- [AltCLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/altclip#transformers.AltCLIPConfig) (AltCLIPConfig 모델)
- **altclip_text_model** -- [AltCLIPTextConfig](/docs/transformers/v5.23.0/ko/model_doc/altclip#transformers.AltCLIPTextConfig) (AltCLIPTextConfig 모델)
- **altclip_vision_model** -- [AltCLIPVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/altclip#transformers.AltCLIPVisionConfig) (AltCLIPVisionConfig 모델)
- **apertus** -- [ApertusConfig](/docs/transformers/v5.23.0/ko/model_doc/apertus#transformers.ApertusConfig) (ApertusConfig 모델)
- **arcee** -- [ArceeConfig](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeConfig) (ArceeConfig 모델)
- **aria** -- [AriaConfig](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaConfig) (AriaConfig 모델)
- **aria_text** -- [AriaTextConfig](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaTextConfig) (AriaTextConfig 모델)
- **audio-spectrogram-transformer** -- [ASTConfig](/docs/transformers/v5.23.0/ko/model_doc/audio-spectrogram-transformer#transformers.ASTConfig) (ASTConfig 모델)
- **audioflamingo3** -- [AudioFlamingo3Config](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3Config) (AudioFlamingo3Config 모델)
- **audioflamingo3_encoder** -- [AudioFlamingo3EncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3EncoderConfig) (AudioFlamingo3EncoderConfig 모델)
- **autoformer** -- [AutoformerConfig](/docs/transformers/v5.23.0/ko/model_doc/autoformer#transformers.AutoformerConfig) (AutoformerConfig 모델)
- **axk1** -- [AXK1Config](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1Config) (AXK1Config 모델)
- **axk2** -- [AXK2Config](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2Config) (AXK2Config 모델)
- **aya_vision** -- [AyaVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/aya_vision#transformers.AyaVisionConfig) (AyaVisionConfig 모델)
- **bailing_hybrid** -- [BailingHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/bailing_hybrid#transformers.BailingHybridConfig) (BailingHybridConfig 모델)
- **bamba** -- [BambaConfig](/docs/transformers/v5.23.0/ko/model_doc/bamba#transformers.BambaConfig) (BambaConfig 모델)
- **bark** -- [BarkConfig](/docs/transformers/v5.23.0/ko/model_doc/bark#transformers.BarkConfig) (BarkConfig 모델)
- **bart** -- [BartConfig](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartConfig) (BartConfig 모델)
- **beam** -- [BeamConfig](/docs/transformers/v5.23.0/ko/model_doc/beam#transformers.BeamConfig) (BeamConfig 모델)
- **beit** -- [BeitConfig](/docs/transformers/v5.23.0/ko/model_doc/beit#transformers.BeitConfig) (BeitConfig 모델)
- **bert** -- [BertConfig](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertConfig) (BertConfig 모델)
- **bert-generation** -- [BertGenerationConfig](/docs/transformers/v5.23.0/ko/model_doc/bert-generation#transformers.BertGenerationConfig) (BertGenerationConfig 모델)
- **big_bird** -- [BigBirdConfig](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdConfig) (BigBirdConfig 모델)
- **bigbird_pegasus** -- [BigBirdPegasusConfig](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusConfig) (BigBirdPegasusConfig 모델)
- **biogpt** -- [BioGptConfig](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptConfig) (BioGptConfig 모델)
- **bit** -- [BitConfig](/docs/transformers/v5.23.0/ko/model_doc/bit#transformers.BitConfig) (BitConfig 모델)
- **bitnet** -- [BitNetConfig](/docs/transformers/v5.23.0/ko/model_doc/bitnet#transformers.BitNetConfig) (BitNetConfig 모델)
- **blenderbot** -- [BlenderbotConfig](/docs/transformers/v5.23.0/ko/model_doc/blenderbot#transformers.BlenderbotConfig) (BlenderbotConfig 모델)
- **blenderbot-small** -- [BlenderbotSmallConfig](/docs/transformers/v5.23.0/ko/model_doc/blenderbot-small#transformers.BlenderbotSmallConfig) (BlenderbotSmallConfig 모델)
- **blip** -- [BlipConfig](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipConfig) (BlipConfig 모델)
- **blip-2** -- [Blip2Config](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2Config) (Blip2Config 모델)
- **blip_2_qformer** -- [Blip2QFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2QFormerConfig) (Blip2QFormerConfig 모델)
- **blip_2_vision_model** -- [Blip2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2VisionConfig) (Blip2VisionConfig 모델)
- **blip_text_model** -- [BlipTextConfig](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipTextConfig) (BlipTextConfig 모델)
- **blip_vision_model** -- [BlipVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipVisionConfig) (BlipVisionConfig 모델)
- **bloom** -- [BloomConfig](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomConfig) (BloomConfig 모델)
- **blt** -- [BltConfig](/docs/transformers/v5.23.0/ko/model_doc/blt#transformers.BltConfig) (BltConfig 모델)
- **blt_global_transformer** -- `BltGlobalTransformerConfig` (BltGlobalTransformerConfig 모델)
- **blt_local_decoder** -- `BltLocalDecoderConfig` (BltLocalDecoderConfig 모델)
- **blt_local_encoder** -- `BltLocalEncoderConfig` (BltLocalEncoderConfig 모델)
- **blt_patcher** -- `BltPatcherConfig` (BltPatcherConfig 모델)
- **bridgetower** -- [BridgeTowerConfig](/docs/transformers/v5.23.0/ko/model_doc/bridgetower#transformers.BridgeTowerConfig) (BridgeTowerConfig 모델)
- **bridgetower_text_model** -- [BridgeTowerTextConfig](/docs/transformers/v5.23.0/ko/model_doc/bridgetower#transformers.BridgeTowerTextConfig) (BridgeTowerTextConfig 모델)
- **bridgetower_vision_model** -- [BridgeTowerVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/bridgetower#transformers.BridgeTowerVisionConfig) (BridgeTowerVisionConfig 모델)
- **bros** -- [BrosConfig](/docs/transformers/v5.23.0/ko/model_doc/bros#transformers.BrosConfig) (BrosConfig 모델)
- **camembert** -- [CamembertConfig](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertConfig) (CamembertConfig 모델)
- **canary** -- [CanaryConfig](/docs/transformers/v5.23.0/ko/model_doc/canary#transformers.CanaryConfig) (CanaryConfig 모델)
- **canary_decoder** -- [CanaryDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/canary#transformers.CanaryDecoderConfig) (CanaryDecoderConfig 모델)
- **canine** -- [CanineConfig](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineConfig) (CanineConfig 모델)
- **chameleon** -- [ChameleonConfig](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonConfig) (ChameleonConfig 모델)
- **chameleon_vqgan** -- [ChameleonVQVAEConfig](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonVQVAEConfig) (ChameleonVQVAEConfig 모델)
- **chinese_clip** -- [ChineseCLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPConfig) (ChineseCLIPConfig 모델)
- **chinese_clip_text_model** -- [ChineseCLIPTextConfig](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPTextConfig) (ChineseCLIPTextConfig 모델)
- **chinese_clip_vision_model** -- [ChineseCLIPVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPVisionConfig) (ChineseCLIPVisionConfig 모델)
- **chmv2** -- [CHMv2Config](/docs/transformers/v5.23.0/ko/model_doc/chmv2#transformers.CHMv2Config) (CHMv2Config 모델)
- **clap** -- [ClapConfig](/docs/transformers/v5.23.0/ko/model_doc/clap#transformers.ClapConfig) (ClapConfig 모델)
- **clap_audio_model** -- [ClapAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/clap#transformers.ClapAudioConfig) (ClapAudioConfig 모델)
- **clap_text_model** -- [ClapTextConfig](/docs/transformers/v5.23.0/ko/model_doc/clap#transformers.ClapTextConfig) (ClapTextConfig 모델)
- **clip** -- [CLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPConfig) (CLIPConfig 모델)
- **clip_text_model** -- [CLIPTextConfig](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTextConfig) (CLIPTextConfig 모델)
- **clip_vision_model** -- [CLIPVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPVisionConfig) (CLIPVisionConfig 모델)
- **clipseg** -- [CLIPSegConfig](/docs/transformers/v5.23.0/ko/model_doc/clipseg#transformers.CLIPSegConfig) (CLIPSegConfig 모델)
- **clipseg_text_model** -- [CLIPSegTextConfig](/docs/transformers/v5.23.0/ko/model_doc/clipseg#transformers.CLIPSegTextConfig) (CLIPSegTextConfig 모델)
- **clipseg_vision_model** -- [CLIPSegVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/clipseg#transformers.CLIPSegVisionConfig) (CLIPSegVisionConfig 모델)
- **clvp** -- [ClvpConfig](/docs/transformers/v5.23.0/ko/model_doc/clvp#transformers.ClvpConfig) (ClvpConfig 모델)
- **clvp_decoder** -- [ClvpDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/clvp#transformers.ClvpDecoderConfig) (ClvpDecoderConfig 모델)
- **clvp_encoder** -- [ClvpEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/clvp#transformers.ClvpEncoderConfig) (ClvpEncoderConfig 모델)
- **codegen** -- [CodeGenConfig](/docs/transformers/v5.23.0/ko/model_doc/codegen#transformers.CodeGenConfig) (CodeGenConfig 모델)
- **cohere** -- [CohereConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere#transformers.CohereConfig) (CohereConfig 모델)
- **cohere2** -- [Cohere2Config](/docs/transformers/v5.23.0/ko/model_doc/cohere2#transformers.Cohere2Config) (Cohere2Config 모델)
- **cohere2_moe** -- [Cohere2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere2_moe#transformers.Cohere2MoeConfig) (Cohere2MoeConfig 모델)
- **cohere2_vision** -- [Cohere2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere2_vision#transformers.Cohere2VisionConfig) (Cohere2VisionConfig 모델)
- **cohere_asr** -- [CohereAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_asr#transformers.CohereAsrConfig) (CohereAsrConfig 모델)
- **cohere_compass** -- [CohereCompassConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassConfig) (CohereCompassConfig 모델)
- **cohere_compass_text** -- [CohereCompassTextConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassTextConfig) (CohereCompassTextConfig 모델)
- **cohere_compass_vision** -- [CohereCompassVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassVisionConfig) (CohereCompassVisionConfig 모델)
- **colmodernvbert** -- [ColModernVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/colmodernvbert#transformers.ColModernVBertConfig) (ColModernVBertConfig 모델)
- **colpali** -- [ColPaliConfig](/docs/transformers/v5.23.0/ko/model_doc/colpali#transformers.ColPaliConfig) (ColPaliConfig 모델)
- **colqwen2** -- [ColQwen2Config](/docs/transformers/v5.23.0/ko/model_doc/colqwen2#transformers.ColQwen2Config) (ColQwen2Config 모델)
- **conditional_detr** -- [ConditionalDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/conditional_detr#transformers.ConditionalDetrConfig) (ConditionalDetrConfig 모델)
- **convbert** -- [ConvBertConfig](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertConfig) (ConvBertConfig 모델)
- **convnext** -- [ConvNextConfig](/docs/transformers/v5.23.0/ko/model_doc/convnext#transformers.ConvNextConfig) (ConvNextConfig 모델)
- **convnextv2** -- [ConvNextV2Config](/docs/transformers/v5.23.0/ko/model_doc/convnextv2#transformers.ConvNextV2Config) (ConvNextV2Config 모델)
- **cosmos3_edge** -- [Cosmos3EdgeConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeConfig) (Cosmos3EdgeConfig 모델)
- **cosmos3_edge_text** -- [Cosmos3EdgeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeTextConfig) (Cosmos3EdgeTextConfig 모델)
- **cosmos3_edge_vision** -- [Cosmos3EdgeVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeVisionConfig) (Cosmos3EdgeVisionConfig 모델)
- **cosmos3_omni** -- [Cosmos3OmniConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_omni#transformers.Cosmos3OmniConfig) (Cosmos3OmniConfig 모델)
- **cpmant** -- [CpmAntConfig](/docs/transformers/v5.23.0/ko/model_doc/cpmant#transformers.CpmAntConfig) (CpmAntConfig 모델)
- **csm** -- [CsmConfig](/docs/transformers/v5.23.0/ko/model_doc/csm#transformers.CsmConfig) (CsmConfig 모델)
- **csm_depth_decoder_model** -- [CsmDepthDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/csm#transformers.CsmDepthDecoderConfig) (CsmDepthDecoderConfig 모델)
- **ctrl** -- [CTRLConfig](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLConfig) (CTRLConfig 모델)
- **cvt** -- [CvtConfig](/docs/transformers/v5.23.0/ko/model_doc/cvt#transformers.CvtConfig) (CvtConfig 모델)
- **cwm** -- [CwmConfig](/docs/transformers/v5.23.0/ko/model_doc/cwm#transformers.CwmConfig) (CwmConfig 모델)
- **d_fine** -- [DFineConfig](/docs/transformers/v5.23.0/ko/model_doc/d_fine#transformers.DFineConfig) (DFineConfig 모델)
- **dab-detr** -- [DabDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/dab-detr#transformers.DabDetrConfig) (DabDetrConfig 모델)
- **dac** -- [DacConfig](/docs/transformers/v5.23.0/ko/model_doc/dac#transformers.DacConfig) (DacConfig 모델)
- **data2vec-audio** -- [Data2VecAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioConfig) (Data2VecAudioConfig 모델)
- **data2vec-text** -- [Data2VecTextConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextConfig) (Data2VecTextConfig 모델)
- **data2vec-vision** -- [Data2VecVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecVisionConfig) (Data2VecVisionConfig 모델)
- **dbrx** -- [DbrxConfig](/docs/transformers/v5.23.0/ko/model_doc/dbrx#transformers.DbrxConfig) (DbrxConfig 모델)
- **deberta** -- [DebertaConfig](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaConfig) (DebertaConfig 모델)
- **deberta-v2** -- [DebertaV2Config](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2Config) (DebertaV2Config 모델)
- **decision_transformer** -- [DecisionTransformerConfig](/docs/transformers/v5.23.0/ko/model_doc/decision_transformer#transformers.DecisionTransformerConfig) (DecisionTransformerConfig 모델)
- **deepseek_ocr2** -- [DeepseekOcr2Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2Config) (DeepseekOcr2Config 모델)
- **deepseek_ocr2_encoder** -- [DeepseekOcr2VisionEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2VisionEncoderConfig) (DeepseekOcr2VisionEncoderConfig 모델)
- **deepseek_ocr2_sam_vision_model** -- [DeepseekOcr2SamVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2SamVisionConfig) (DeepseekOcr2SamVisionConfig 모델)
- **deepseek_ocr2_text** -- [DeepseekOcr2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2TextConfig) (DeepseekOcr2TextConfig 모델)
- **deepseek_ocr2_vision** -- [DeepseekOcr2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2VisionConfig) (DeepseekOcr2VisionConfig 모델)
- **deepseek_v2** -- [DeepseekV2Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v2#transformers.DeepseekV2Config) (DeepseekV2Config 모델)
- **deepseek_v3** -- [DeepseekV3Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3Config) (DeepseekV3Config 모델)
- **deepseek_v32** -- [DeepseekV32Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v32#transformers.DeepseekV32Config) (DeepseekV32Config 모델)
- **deepseek_v4** -- [DeepseekV4Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v4#transformers.DeepseekV4Config) (DeepseekV4Config 모델)
- **deepseek_vl** -- [DeepseekVLConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl#transformers.DeepseekVLConfig) (DeepseekVLConfig 모델)
- **deepseek_vl_hybrid** -- [DeepseekVLHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl_hybrid#transformers.DeepseekVLHybridConfig) (DeepseekVLHybridConfig 모델)
- **deformable_detr** -- [DeformableDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/deformable_detr#transformers.DeformableDetrConfig) (DeformableDetrConfig 모델)
- **deimv2** -- [Deimv2Config](/docs/transformers/v5.23.0/ko/model_doc/deimv2#transformers.Deimv2Config) (Deimv2Config 모델)
- **deit** -- [DeiTConfig](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTConfig) (DeiTConfig 모델)
- **depth_anything** -- [DepthAnythingConfig](/docs/transformers/v5.23.0/ko/model_doc/depth_anything#transformers.DepthAnythingConfig) (DepthAnythingConfig 모델)
- **depth_pro** -- [DepthProConfig](/docs/transformers/v5.23.0/ko/model_doc/depth_pro#transformers.DepthProConfig) (DepthProConfig 모델)
- **detr** -- [DetrConfig](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrConfig) (DetrConfig 모델)
- **dia** -- [DiaConfig](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaConfig) (DiaConfig 모델)
- **dia_decoder** -- [DiaDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaDecoderConfig) (DiaDecoderConfig 모델)
- **dia_encoder** -- [DiaEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaEncoderConfig) (DiaEncoderConfig 모델)
- **diffllama** -- [DiffLlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaConfig) (DiffLlamaConfig 모델)
- **diffusion_gemma** -- [DiffusionGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/diffusion_gemma#transformers.DiffusionGemmaConfig) (DiffusionGemmaConfig 모델)
- **diffusion_gemma_text** -- [DiffusionGemmaTextConfig](/docs/transformers/v5.23.0/ko/model_doc/diffusion_gemma#transformers.DiffusionGemmaTextConfig) (DiffusionGemmaTextConfig 모델)
- **dinat** -- [DinatConfig](/docs/transformers/v5.23.0/ko/model_doc/dinat#transformers.DinatConfig) (DinatConfig 모델)
- **dinov2** -- [Dinov2Config](/docs/transformers/v5.23.0/ko/model_doc/dinov2#transformers.Dinov2Config) (Dinov2Config 모델)
- **dinov2_with_registers** -- [Dinov2WithRegistersConfig](/docs/transformers/v5.23.0/ko/model_doc/dinov2_with_registers#transformers.Dinov2WithRegistersConfig) (Dinov2WithRegistersConfig 모델)
- **dinov3_convnext** -- [DINOv3ConvNextConfig](/docs/transformers/v5.23.0/ko/model_doc/dinov3#transformers.DINOv3ConvNextConfig) (DINOv3ConvNextConfig 모델)
- **dinov3_vit** -- [DINOv3ViTConfig](/docs/transformers/v5.23.0/ko/model_doc/dinov3#transformers.DINOv3ViTConfig) (DINOv3ViTConfig 모델)
- **distilbert** -- [DistilBertConfig](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertConfig) (DistilBertConfig 모델)
- **doge** -- [DogeConfig](/docs/transformers/v5.23.0/ko/model_doc/doge#transformers.DogeConfig) (DogeConfig 모델)
- **donut-swin** -- [DonutSwinConfig](/docs/transformers/v5.23.0/ko/model_doc/donut#transformers.DonutSwinConfig) (DonutSwinConfig 모델)
- **dots1** -- [Dots1Config](/docs/transformers/v5.23.0/ko/model_doc/dots1#transformers.Dots1Config) (Dots1Config 모델)
- **dots3_note** -- [Dots3NoteConfig](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteConfig) (Dots3NoteConfig 모델)
- **dots3_note_audio_encoder** -- [Dots3NoteAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteAudioConfig) (Dots3NoteAudioConfig 모델)
- **dots3_note_text** -- [Dots3NoteTextConfig](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteTextConfig) (Dots3NoteTextConfig 모델)
- **dots3_note_vision_encoder** -- [Dots3NoteVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteVisionConfig) (Dots3NoteVisionConfig 모델)
- **dpr** -- [DPRConfig](/docs/transformers/v5.23.0/ko/model_doc/dpr#transformers.DPRConfig) (DPRConfig 모델)
- **dpt** -- [DPTConfig](/docs/transformers/v5.23.0/ko/model_doc/dpt#transformers.DPTConfig) (DPTConfig 모델)
- **edgetam** -- [EdgeTamConfig](/docs/transformers/v5.23.0/ko/model_doc/edgetam#transformers.EdgeTamConfig) (EdgeTamConfig 모델)
- **edgetam_mask_decoder** -- [EdgeTamMaskDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/edgetam#transformers.EdgeTamMaskDecoderConfig) (EdgeTamMaskDecoderConfig 모델)
- **edgetam_prompt_encoder** -- [EdgeTamPromptEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/edgetam#transformers.EdgeTamPromptEncoderConfig) (EdgeTamPromptEncoderConfig 모델)
- **edgetam_video** -- [EdgeTamVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/edgetam_video#transformers.EdgeTamVideoConfig) (EdgeTamVideoConfig 모델)
- **edgetam_video_mask_decoder** -- [EdgeTamVideoMaskDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/edgetam_video#transformers.EdgeTamVideoMaskDecoderConfig) (EdgeTamVideoMaskDecoderConfig 모델)
- **edgetam_video_prompt_encoder** -- [EdgeTamVideoPromptEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/edgetam_video#transformers.EdgeTamVideoPromptEncoderConfig) (EdgeTamVideoPromptEncoderConfig 모델)
- **edgetam_vision_model** -- [EdgeTamVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/edgetam#transformers.EdgeTamVisionConfig) (EdgeTamVisionConfig 모델)
- **efficientloftr** -- [EfficientLoFTRConfig](/docs/transformers/v5.23.0/ko/model_doc/efficientloftr#transformers.EfficientLoFTRConfig) (EfficientLoFTRConfig 모델)
- **efficientnet** -- [EfficientNetConfig](/docs/transformers/v5.23.0/ko/model_doc/efficientnet#transformers.EfficientNetConfig) (EfficientNetConfig 모델)
- **electra** -- [ElectraConfig](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraConfig) (ElectraConfig 모델)
- **embedding_gemma2** -- [EmbeddingGemma2Config](/docs/transformers/v5.23.0/ko/model_doc/embedding_gemma2#transformers.EmbeddingGemma2Config) (EmbeddingGemma2Config 모델)
- **embedding_gemma2_text** -- [EmbeddingGemma2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/embedding_gemma2#transformers.EmbeddingGemma2TextConfig) (EmbeddingGemma2TextConfig 모델)
- **emu3** -- [Emu3Config](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3Config) (Emu3Config 모델)
- **emu3_text_model** -- [Emu3TextConfig](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3TextConfig) (Emu3TextConfig 모델)
- **emu3_vqgan** -- [Emu3VQVAEConfig](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3VQVAEConfig) (Emu3VQVAEConfig 모델)
- **encodec** -- [EncodecConfig](/docs/transformers/v5.23.0/ko/model_doc/encodec#transformers.EncodecConfig) (EncodecConfig 모델)
- **인코더-디코더** -- [EncoderDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/encoder-decoder#transformers.EncoderDecoderConfig) (EncoderDecoderConfig 모델)
- **eomt** -- [EomtConfig](/docs/transformers/v5.23.0/ko/model_doc/eomt#transformers.EomtConfig) (EomtConfig 모델)
- **eomt_dinov3** -- [EomtDinov3Config](/docs/transformers/v5.23.0/ko/model_doc/eomt_dinov3#transformers.EomtDinov3Config) (EomtDinov3Config 모델)
- **ernie** -- [ErnieConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieConfig) (ErnieConfig 모델)
- **ernie4_5** -- [Ernie4_5Config](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5#transformers.Ernie4_5Config) (Ernie4_5Config 모델)
- **ernie4_5_moe** -- [Ernie4_5_MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_moe#transformers.Ernie4_5_MoeConfig) (Ernie4_5_MoeConfig 모델)
- **ernie4_5_vl_moe** -- [Ernie4_5_VLMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeConfig) (Ernie4_5_VLMoeConfig 모델)
- **ernie4_5_vl_moe_text** -- [Ernie4_5_VLMoeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeTextConfig) (Ernie4_5_VLMoeTextConfig 모델)
- **ernie4_5_vl_moe_vision** -- [Ernie4_5_VLMoeVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeVisionConfig) (Ernie4_5_VLMoeVisionConfig 모델)
- **esm** -- [EsmConfig](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmConfig) (EsmConfig 모델)
- **esmc** -- [EsmcConfig](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcConfig) (EsmcConfig 모델)
- **esmfold2** -- [EsmFold2Config](/docs/transformers/v5.23.0/ko/model_doc/esmfold2#transformers.EsmFold2Config) (EsmFold2Config 모델)
- **eurobert** -- [EuroBertConfig](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertConfig) (EuroBertConfig 모델)
- **evolla** -- [EvollaConfig](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaConfig) (EvollaConfig 모델)
- **exaone4** -- [Exaone4Config](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4Config) (Exaone4Config 모델)
- **exaone4_5** -- [Exaone4_5_Config](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_Config) (Exaone4_5_Config 모델)
- **exaone4_5_vision** -- [Exaone4_5_VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_VisionConfig) (Exaone4_5_VisionConfig 모델)
- **exaone_moe** -- [ExaoneMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/exaone_moe#transformers.ExaoneMoeConfig) (ExaoneMoeConfig 모델)
- **falcon** -- [FalconConfig](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconConfig) (FalconConfig 모델)
- **falcon_h1** -- [FalconH1Config](/docs/transformers/v5.23.0/ko/model_doc/falcon_h1#transformers.FalconH1Config) (FalconH1Config 모델)
- **falcon_mamba** -- [FalconMambaConfig](/docs/transformers/v5.23.0/ko/model_doc/falcon_mamba#transformers.FalconMambaConfig) (FalconMambaConfig 모델)
- **fast_vlm** -- [FastVlmConfig](/docs/transformers/v5.23.0/ko/model_doc/fast_vlm#transformers.FastVlmConfig) (FastVlmConfig 모델)
- **fastspeech2_conformer** -- [FastSpeech2ConformerConfig](/docs/transformers/v5.23.0/ko/model_doc/fastspeech2_conformer#transformers.FastSpeech2ConformerConfig) (FastSpeech2ConformerConfig 모델)
- **fastspeech2_conformer_hifigan** -- [FastSpeech2ConformerHifiGanConfig](/docs/transformers/v5.23.0/ko/model_doc/fastspeech2_conformer#transformers.FastSpeech2ConformerHifiGanConfig) (FastSpeech2ConformerHifiGanConfig 모델)
- **fastspeech2_conformer_with_hifigan** -- [FastSpeech2ConformerWithHifiGanConfig](/docs/transformers/v5.23.0/ko/model_doc/fastspeech2_conformer#transformers.FastSpeech2ConformerWithHifiGanConfig) (FastSpeech2ConformerWithHifiGanConfig 모델)
- **flaubert** -- [FlaubertConfig](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertConfig) (FlaubertConfig 모델)
- **flava** -- [FlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaConfig) (FlavaConfig 모델)
- **flava_image_codebook** -- [FlavaImageCodebookConfig](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaImageCodebookConfig) (FlavaImageCodebookConfig 모델)
- **flava_image_model** -- [FlavaImageConfig](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaImageConfig) (FlavaImageConfig 모델)
- **flava_multimodal_model** -- [FlavaMultimodalConfig](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaMultimodalConfig) (FlavaMultimodalConfig 모델)
- **flava_text_model** -- [FlavaTextConfig](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaTextConfig) (FlavaTextConfig 모델)
- **flex_olmo** -- [FlexOlmoConfig](/docs/transformers/v5.23.0/ko/model_doc/flex_olmo#transformers.FlexOlmoConfig) (FlexOlmoConfig 모델)
- **flint** -- [FlintConfig](/docs/transformers/v5.23.0/ko/model_doc/flint#transformers.FlintConfig) (FlintConfig 모델)
- **florence2** -- [Florence2Config](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2Config) (Florence2Config 모델)
- **florence_vision** -- [Florence2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2VisionConfig) (Florence2VisionConfig 모델)
- **fnet** -- [FNetConfig](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetConfig) (FNetConfig 모델)
- **focalnet** -- [FocalNetConfig](/docs/transformers/v5.23.0/ko/model_doc/focalnet#transformers.FocalNetConfig) (FocalNetConfig 모델)
- **fsmt** -- [FSMTConfig](/docs/transformers/v5.23.0/ko/model_doc/fsmt#transformers.FSMTConfig) (FSMTConfig 모델)
- **fun_asr_nano** -- [FunAsrNanoConfig](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoConfig) (FunAsrNanoConfig 모델)
- **fun_asr_nano_encoder** -- [FunAsrNanoEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoEncoderConfig) (FunAsrNanoEncoderConfig 모델)
- **funnel** -- [FunnelConfig](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelConfig) (FunnelConfig 모델)
- **fuyu** -- [FuyuConfig](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuConfig) (FuyuConfig 모델)
- **gemma** -- [GemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaConfig) (GemmaConfig 모델)
- **gemma2** -- [Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2Config) (Gemma2Config 모델)
- **gemma3** -- [Gemma3Config](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Config) (Gemma3Config 모델)
- **gemma3_text** -- [Gemma3TextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3TextConfig) (Gemma3TextConfig 모델)
- **gemma3n** -- [Gemma3nConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nConfig) (Gemma3nConfig 모델)
- **gemma3n_audio** -- [Gemma3nAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nAudioConfig) (Gemma3nAudioConfig 모델)
- **gemma3n_text** -- [Gemma3nTextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nTextConfig) (Gemma3nTextConfig 모델)
- **gemma3n_vision** -- [Gemma3nVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nVisionConfig) (Gemma3nVisionConfig 모델)
- **gemma4** -- [Gemma4Config](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4Config) (Gemma4Config 모델)
- **gemma4_assistant** -- [Gemma4AssistantConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_assistant#transformers.Gemma4AssistantConfig) (Gemma4AssistantConfig 모델)
- **gemma4_audio** -- [Gemma4AudioConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4AudioConfig) (Gemma4AudioConfig 모델)
- **gemma4_text** -- [Gemma4TextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4TextConfig) (Gemma4TextConfig 모델)
- **gemma4_unified** -- [Gemma4UnifiedConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedConfig) (Gemma4UnifiedConfig 모델)
- **gemma4_unified_assistant** -- [Gemma4UnifiedAssistantConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified_assistant#transformers.Gemma4UnifiedAssistantConfig) (Gemma4UnifiedAssistantConfig 모델)
- **gemma4_unified_audio** -- [Gemma4UnifiedAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedAudioConfig) (Gemma4UnifiedAudioConfig 모델)
- **gemma4_unified_text** -- [Gemma4UnifiedTextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedTextConfig) (Gemma4UnifiedTextConfig 모델)
- **gemma4_unified_vision** -- [Gemma4UnifiedVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedVisionConfig) (Gemma4UnifiedVisionConfig 모델)
- **gemma4_vision** -- [Gemma4VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4VisionConfig) (Gemma4VisionConfig 모델)
- **git** -- [GitConfig](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitConfig) (GitConfig 모델)
- **git_vision_model** -- [GitVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitVisionConfig) (GitVisionConfig 모델)
- **glm** -- [GlmConfig](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmConfig) (GlmConfig 모델)
- **glm4** -- [Glm4Config](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4Config) (Glm4Config 모델)
- **glm46v** -- [Glm46VConfig](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VConfig) (Glm46VConfig 모델)
- **glm4_moe** -- [Glm4MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe#transformers.Glm4MoeConfig) (Glm4MoeConfig 모델)
- **glm4_moe_lite** -- [Glm4MoeLiteConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe_lite#transformers.Glm4MoeLiteConfig) (Glm4MoeLiteConfig 모델)
- **glm4v** -- [Glm4vConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vConfig) (Glm4vConfig 모델)
- **glm4v_moe** -- [Glm4vMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeConfig) (Glm4vMoeConfig 모델)
- **glm4v_moe_text** -- [Glm4vMoeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeTextConfig) (Glm4vMoeTextConfig 모델)
- **glm4v_moe_vision** -- [Glm4vMoeVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeVisionConfig) (Glm4vMoeVisionConfig 모델)
- **glm4v_text** -- [Glm4vTextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vTextConfig) (Glm4vTextConfig 모델)
- **glm4v_vision** -- [Glm4vVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vVisionConfig) (Glm4vVisionConfig 모델)
- **glm5_next** -- [Glm5NextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextConfig) (Glm5NextConfig 모델)
- **glm5_next_text** -- [Glm5NextTextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextTextConfig) (Glm5NextTextConfig 모델)
- **glm5_next_vision** -- [Glm5NextVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextVisionConfig) (Glm5NextVisionConfig 모델)
- **glm_image** -- [GlmImageConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageConfig) (GlmImageConfig 모델)
- **glm_image_text** -- [GlmImageTextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageTextConfig) (GlmImageTextConfig 모델)
- **glm_image_vision** -- [GlmImageVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageVisionConfig) (GlmImageVisionConfig 모델)
- **glm_image_vqmodel** -- [GlmImageVQVAEConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageVQVAEConfig) (GlmImageVQVAEConfig 모델)
- **glm_moe_dsa** -- [GlmMoeDsaConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_moe_dsa#transformers.GlmMoeDsaConfig) (GlmMoeDsaConfig 모델)
- **glm_ocr** -- [GlmOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrConfig) (GlmOcrConfig 모델)
- **glm_ocr_text** -- [GlmOcrTextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrTextConfig) (GlmOcrTextConfig 모델)
- **glm_ocr_vision** -- [GlmOcrVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrVisionConfig) (GlmOcrVisionConfig 모델)
- **glmasr** -- [GlmAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrConfig) (GlmAsrConfig 모델)
- **glmasr_encoder** -- [GlmAsrEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrEncoderConfig) (GlmAsrEncoderConfig 모델)
- **glmga** -- [GlmgaConfig](/docs/transformers/v5.23.0/ko/model_doc/glmga#transformers.GlmgaConfig) (GlmgaConfig 모델)
- **glpn** -- [GLPNConfig](/docs/transformers/v5.23.0/ko/model_doc/glpn#transformers.GLPNConfig) (GLPNConfig 모델)
- **got_ocr2** -- [GotOcr2Config](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2Config) (GotOcr2Config 모델)
- **got_ocr2_vision** -- [GotOcr2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2VisionConfig) (GotOcr2VisionConfig 모델)
- **gpt-sw3** -- [GPT2Config](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Config) (GPT2Config 모델)
- **gpt2** -- [GPT2Config](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Config) (GPT2Config 모델)
- **gpt_bigcode** -- [GPTBigCodeConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeConfig) (GPTBigCodeConfig 모델)
- **gpt_neo** -- [GPTNeoConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoConfig) (GPTNeoConfig 모델)
- **gpt_neox** -- [GPTNeoXConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXConfig) (GPTNeoXConfig 모델)
- **gpt_neox_japanese** -- [GPTNeoXJapaneseConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox_japanese#transformers.GPTNeoXJapaneseConfig) (GPTNeoXJapaneseConfig 모델)
- **gpt_oss** -- [GptOssConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssConfig) (GptOssConfig 모델)
- **gptj** -- [GPTJConfig](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJConfig) (GPTJConfig 모델)
- **gradium_phonon_tts** -- [GradiumPhononTTSConfig](/docs/transformers/v5.23.0/ko/model_doc/gradium_phonon_tts#transformers.GradiumPhononTTSConfig) (GradiumPhononTTSConfig 모델)
- **granite** -- [GraniteConfig](/docs/transformers/v5.23.0/ko/model_doc/granite#transformers.GraniteConfig) (GraniteConfig 모델)
- **granite4_vision** -- [Granite4VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionConfig) (Granite4VisionConfig 모델)
- **granite4_vision_text** -- [Granite4VisionTextConfig](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionTextConfig) (Granite4VisionTextConfig 모델)
- **granite_docling** -- [GraniteDoclingConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingConfig) (GraniteDoclingConfig 모델)
- **granite_docling_text** -- [GraniteDoclingTextConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingTextConfig) (GraniteDoclingTextConfig 모델)
- **granite_docling_vision** -- [GraniteDoclingVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingVisionConfig) (GraniteDoclingVisionConfig 모델)
- **granite_speech** -- [GraniteSpeechConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechConfig) (GraniteSpeechConfig 모델)
- **granite_speech5_ctc** -- [GraniteSpeech5CTCConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5CTCConfig) (GraniteSpeech5CTCConfig 모델)
- **granite_speech5_encoder** -- [GraniteSpeech5EncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5EncoderConfig) (GraniteSpeech5EncoderConfig 모델)
- **granite_speech_encoder** -- [GraniteSpeechEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechEncoderConfig) (GraniteSpeechEncoderConfig 모델)
- **granite_speech_plus** -- [GraniteSpeechPlusConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusConfig) (GraniteSpeechPlusConfig 모델)
- **granite_speech_plus_encoder** -- [GraniteSpeechPlusEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusEncoderConfig) (GraniteSpeechPlusEncoderConfig 모델)
- **granite_swa** -- [GraniteSWAConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_swa#transformers.GraniteSWAConfig) (GraniteSWAConfig 모델)
- **granitemoe** -- [GraniteMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoe#transformers.GraniteMoeConfig) (GraniteMoeConfig 모델)
- **granitemoe_swa** -- [GraniteMoeSWAConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoe_swa#transformers.GraniteMoeSWAConfig) (GraniteMoeSWAConfig 모델)
- **granitemoehybrid** -- [GraniteMoeHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoehybrid#transformers.GraniteMoeHybridConfig) (GraniteMoeHybridConfig 모델)
- **granitemoeshared** -- [GraniteMoeSharedConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoeshared#transformers.GraniteMoeSharedConfig) (GraniteMoeSharedConfig 모델)
- **grounding-dino** -- [GroundingDinoConfig](/docs/transformers/v5.23.0/ko/model_doc/grounding-dino#transformers.GroundingDinoConfig) (GroundingDinoConfig 모델)
- **groupvit** -- [GroupViTConfig](/docs/transformers/v5.23.0/ko/model_doc/groupvit#transformers.GroupViTConfig) (GroupViTConfig 모델)
- **groupvit_text_model** -- [GroupViTTextConfig](/docs/transformers/v5.23.0/ko/model_doc/groupvit#transformers.GroupViTTextConfig) (GroupViTTextConfig 모델)
- **groupvit_vision_model** -- [GroupViTVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/groupvit#transformers.GroupViTVisionConfig) (GroupViTVisionConfig 모델)
- **gte** -- [GteConfig](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteConfig) (GteConfig 모델)
- **helium** -- [HeliumConfig](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumConfig) (HeliumConfig 모델)
- **hgnet_v2** -- [HGNetV2Config](/docs/transformers/v5.23.0/ko/model_doc/hgnet_v2#transformers.HGNetV2Config) (HGNetV2Config 모델)
- **hiera** -- [HieraConfig](/docs/transformers/v5.23.0/ko/model_doc/hiera#transformers.HieraConfig) (HieraConfig 모델)
- **higgs_audio_v2** -- [HiggsAudioV2Config](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2#transformers.HiggsAudioV2Config) (HiggsAudioV2Config 모델)
- **higgs_audio_v2_tokenizer** -- [HiggsAudioV2TokenizerConfig](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2_tokenizer#transformers.HiggsAudioV2TokenizerConfig) (HiggsAudioV2TokenizerConfig 모델)
- **hrm_text** -- [HrmTextConfig](/docs/transformers/v5.23.0/ko/model_doc/hrm_text#transformers.HrmTextConfig) (HrmTextConfig 모델)
- **hubert** -- [HubertConfig](/docs/transformers/v5.23.0/ko/model_doc/hubert#transformers.HubertConfig) (HubertConfig 모델)
- **hunyuan_v1_dense** -- [HunYuanDenseV1Config](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_dense#transformers.HunYuanDenseV1Config) (HunYuanDenseV1Config 모델)
- **hunyuan_v1_moe** -- [HunYuanMoEV1Config](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_moe#transformers.HunYuanMoEV1Config) (HunYuanMoEV1Config 모델)
- **hunyuan_vl** -- [HunYuanVLConfig](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLConfig) (HunYuanVLConfig 모델)
- **hunyuan_vl_text** -- [HunYuanVLTextConfig](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLTextConfig) (HunYuanVLTextConfig 모델)
- **hunyuan_vl_vision** -- [HunYuanVLVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLVisionConfig) (HunYuanVLVisionConfig 모델)
- **hy_v3** -- [HYV3Config](/docs/transformers/v5.23.0/ko/model_doc/hy_v3#transformers.HYV3Config) (HYV3Config 모델)
- **hy_v4** -- [HYV4Config](/docs/transformers/v5.23.0/ko/model_doc/hy_v4#transformers.HYV4Config) (HYV4Config 모델)
- **hyperclovax** -- [HyperCLOVAXConfig](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax#transformers.HyperCLOVAXConfig) (HyperCLOVAXConfig 모델)
- **hyperclovax_vision_v2** -- [HyperCLOVAXVisionV2Config](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax_vision_v2#transformers.HyperCLOVAXVisionV2Config) (HyperCLOVAXVisionV2Config 모델)
- **ibert** -- [IBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertConfig) (IBertConfig 모델)
- **idefics** -- [IdeficsConfig](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsConfig) (IdeficsConfig 모델)
- **idefics2** -- [Idefics2Config](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2Config) (Idefics2Config 모델)
- **idefics2_perceiver** -- [Idefics2PerceiverConfig](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2PerceiverConfig) (Idefics2PerceiverConfig 모델)
- **idefics2_vision** -- [Idefics2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2VisionConfig) (Idefics2VisionConfig 모델)
- **idefics3** -- [Idefics3Config](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3Config) (Idefics3Config 모델)
- **idefics3_vision** -- [Idefics3VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3VisionConfig) (Idefics3VisionConfig 모델)
- **idefics_perciever** -- [IdeficsPerceiverConfig](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsPerceiverConfig) (IdeficsPerceiverConfig 모델)
- **idefics_vision** -- [IdeficsVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsVisionConfig) (IdeficsVisionConfig 모델)
- **ijepa** -- [IJepaConfig](/docs/transformers/v5.23.0/ko/model_doc/ijepa#transformers.IJepaConfig) (IJepaConfig 모델)
- **imagegpt** -- [ImageGPTConfig](/docs/transformers/v5.23.0/ko/model_doc/imagegpt#transformers.ImageGPTConfig) (ImageGPTConfig 모델)
- **informer** -- [InformerConfig](/docs/transformers/v5.23.0/ko/model_doc/informer#transformers.InformerConfig) (InformerConfig 모델)
- **inkling2** -- [Inkling2Config](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2Config) (Inkling2Config 모델)
- **inkling2_text** -- [Inkling2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2TextConfig) (Inkling2TextConfig 모델)
- **inkling2_vision** -- [Inkling2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2VisionConfig) (Inkling2VisionConfig 모델)
- **inkling_audio** -- [InklingAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingAudioConfig) (InklingAudioConfig 모델)
- **inkling_mm_model** -- [InklingConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingConfig) (InklingConfig 모델)
- **inkling_text** -- [InklingTextConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingTextConfig) (InklingTextConfig 모델)
- **inkling_vision** -- [InklingVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingVisionConfig) (InklingVisionConfig 모델)
- **instructblip** -- [InstructBlipConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipConfig) (InstructBlipConfig 모델)
- **instructblip_qformer** -- [InstructBlipQFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipQFormerConfig) (InstructBlipQFormerConfig 모델)
- **instructblip_vision_model** -- [InstructBlipVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipVisionConfig) (InstructBlipVisionConfig 모델)
- **instructblipvideo** -- [InstructBlipVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoConfig) (InstructBlipVideoConfig 모델)
- **instructblipvideo_qformer** -- [InstructBlipVideoQFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoQFormerConfig) (InstructBlipVideoQFormerConfig 모델)
- **instructblipvideo_vision_model** -- [InstructBlipVideoVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoVisionConfig) (InstructBlipVideoVisionConfig 모델)
- **internlm3** -- [InternLM3Config](/docs/transformers/v5.23.0/ko/model_doc/internlm3#transformers.InternLM3Config) (InternLM3Config 모델)
- **internvl** -- [InternVLConfig](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLConfig) (InternVLConfig 모델)
- **internvl_vision** -- [InternVLVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLVisionConfig) (InternVLVisionConfig 모델)
- **iquestcoder** -- [IQuestCoderConfig](/docs/transformers/v5.23.0/ko/model_doc/iquestcoder#transformers.IQuestCoderConfig) (IQuestCoderConfig 모델)
- **jais2** -- [Jais2Config](/docs/transformers/v5.23.0/ko/model_doc/jais2#transformers.Jais2Config) (Jais2Config 모델)
- **jamba** -- [JambaConfig](/docs/transformers/v5.23.0/ko/model_doc/jamba#transformers.JambaConfig) (JambaConfig 모델)
- **janus** -- [JanusConfig](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusConfig) (JanusConfig 모델)
- **janus_vision_model** -- [JanusVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusVisionConfig) (JanusVisionConfig 모델)
- **janus_vqgan** -- [JanusVQVAEConfig](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusVQVAEConfig) (JanusVQVAEConfig 모델)
- **jetmoe** -- [JetMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/jetmoe#transformers.JetMoeConfig) (JetMoeConfig 모델)
- **jev** -- [JevConfig](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevConfig) (JevConfig 모델)
- **jina_embeddings_v3** -- [JinaEmbeddingsV3Config](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3Config) (JinaEmbeddingsV3Config 모델)
- **kimi_k25** -- [Kimi_K25Config](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25Config) (Kimi_K25Config 모델)
- **kimi_k25_vision** -- [Kimi_K25VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25VisionConfig) (Kimi_K25VisionConfig 모델)
- **kimi_linear** -- [KimiLinearConfig](/docs/transformers/v5.23.0/ko/model_doc/kimi_linear#transformers.KimiLinearConfig) (KimiLinearConfig 모델)
- **kolibri1** -- [Kolibri1Config](/docs/transformers/v5.23.0/ko/model_doc/kolibri1#transformers.Kolibri1Config) (Kolibri1Config 모델)
- **kosmos-2** -- [Kosmos2Config](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2Config) (Kosmos2Config 모델)
- **kosmos-2.5** -- [Kosmos2_5Config](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5Config) (Kosmos2_5Config 모델)
- **kosmos_2_5_text_model** -- [Kosmos2_5TextConfig](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5TextConfig) (Kosmos2_5TextConfig 모델)
- **kosmos_2_5_vision_model** -- [Kosmos2_5VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5VisionConfig) (Kosmos2_5VisionConfig 모델)
- **kosmos_2_text_model** -- [Kosmos2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2TextConfig) (Kosmos2TextConfig 모델)
- **kosmos_2_vision_model** -- [Kosmos2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2VisionConfig) (Kosmos2VisionConfig 모델)
- **kyutai_speech_to_text** -- [KyutaiSpeechToTextConfig](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextConfig) (KyutaiSpeechToTextConfig 모델)
- **laguna** -- [LagunaConfig](/docs/transformers/v5.23.0/ko/model_doc/laguna#transformers.LagunaConfig) (LagunaConfig 모델)
- **lasr_ctc** -- [LasrCTCConfig](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrCTCConfig) (LasrCTCConfig 모델)
- **lasr_encoder** -- [LasrEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrEncoderConfig) (LasrEncoderConfig 모델)
- **layoutlm** -- [LayoutLMConfig](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMConfig) (LayoutLMConfig 모델)
- **layoutlmv2** -- [LayoutLMv2Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2Config) (LayoutLMv2Config 모델)
- **layoutlmv3** -- [LayoutLMv3Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3Config) (LayoutLMv3Config 모델)
- **layoutxlm** -- [LayoutXLMConfig](/docs/transformers/v5.23.0/ko/model_doc/layoutxlm#transformers.LayoutXLMConfig) (LayoutXLMConfig 모델)
- **led** -- [LEDConfig](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.LEDConfig) (LEDConfig 모델)
- **levit** -- [LevitConfig](/docs/transformers/v5.23.0/ko/model_doc/levit#transformers.LevitConfig) (LevitConfig 모델)
- **lfm2** -- [Lfm2Config](/docs/transformers/v5.23.0/ko/model_doc/lfm2#transformers.Lfm2Config) (Lfm2Config 모델)
- **lfm2_moe** -- [Lfm2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm2_moe#transformers.Lfm2MoeConfig) (Lfm2MoeConfig 모델)
- **lfm2_vl** -- [Lfm2VlConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm2_vl#transformers.Lfm2VlConfig) (Lfm2VlConfig 모델)
- **lfm3** -- [Lfm3Config](/docs/transformers/v5.23.0/ko/model_doc/lfm3#transformers.Lfm3Config) (Lfm3Config 모델)
- **lfm3_vl** -- [Lfm3VlConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlConfig) (Lfm3VlConfig 모델)
- **lfm3_vl_vision** -- [Lfm3VlVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlVisionConfig) (Lfm3VlVisionConfig 모델)
- **lightglue** -- [LightGlueConfig](/docs/transformers/v5.23.0/ko/model_doc/lightglue#transformers.LightGlueConfig) (LightGlueConfig 모델)
- **lighton_ocr** -- [LightOnOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/lighton_ocr#transformers.LightOnOcrConfig) (LightOnOcrConfig 모델)
- **lilt** -- [LiltConfig](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltConfig) (LiltConfig 모델)
- **llama** -- [LlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaConfig) (LlamaConfig 모델)
- **llama4** -- [Llama4Config](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4Config) (Llama4Config 모델)
- **llama4_text** -- [Llama4TextConfig](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4TextConfig) (Llama4TextConfig 모델)
- **llama4_vision_model** -- [Llama4VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4VisionConfig) (Llama4VisionConfig 모델)
- **llava** -- [LlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaConfig) (LlavaConfig 모델)
- **llava_next** -- [LlavaNextConfig](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextConfig) (LlavaNextConfig 모델)
- **llava_next_video** -- [LlavaNextVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoConfig) (LlavaNextVideoConfig 모델)
- **llava_onevision** -- [LlavaOnevisionConfig](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionConfig) (LlavaOnevisionConfig 모델)
- **longcat_flash** -- [LongcatFlashConfig](/docs/transformers/v5.23.0/ko/model_doc/longcat_flash#transformers.LongcatFlashConfig) (LongcatFlashConfig 모델)
- **longformer** -- [LongformerConfig](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerConfig) (LongformerConfig 모델)
- **longt5** -- [LongT5Config](/docs/transformers/v5.23.0/ko/model_doc/longt5#transformers.LongT5Config) (LongT5Config 모델)
- **luke** -- [LukeConfig](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeConfig) (LukeConfig 모델)
- **lw_detr** -- [LwDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/lw_detr#transformers.LwDetrConfig) (LwDetrConfig 모델)
- **lw_detr_vit** -- [LwDetrViTConfig](/docs/transformers/v5.23.0/ko/model_doc/lw_detr#transformers.LwDetrViTConfig) (LwDetrViTConfig 모델)
- **lxmert** -- [LxmertConfig](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.LxmertConfig) (LxmertConfig 모델)
- **m2m_100** -- [M2M100Config](/docs/transformers/v5.23.0/ko/model_doc/m2m_100#transformers.M2M100Config) (M2M100Config 모델)
- **mamba** -- [MambaConfig](/docs/transformers/v5.23.0/ko/model_doc/mamba#transformers.MambaConfig) (MambaConfig 모델)
- **mamba2** -- [Mamba2Config](/docs/transformers/v5.23.0/ko/model_doc/mamba2#transformers.Mamba2Config) (Mamba2Config 모델)
- **marian** -- [MarianConfig](/docs/transformers/v5.23.0/ko/model_doc/marian#transformers.MarianConfig) (MarianConfig 모델)
- **markuplm** -- [MarkupLMConfig](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMConfig) (MarkupLMConfig 모델)
- **mask2former** -- [Mask2FormerConfig](/docs/transformers/v5.23.0/ko/model_doc/mask2former#transformers.Mask2FormerConfig) (Mask2FormerConfig 모델)
- **maskformer** -- [MaskFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/maskformer#transformers.MaskFormerConfig) (MaskFormerConfig 모델)
- **maskformer-swin** -- `MaskFormerSwinConfig` (MaskFormerSwinConfig 모델)
- **mbart** -- [MBartConfig](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartConfig) (MBartConfig 모델)
- **megatron-bert** -- [MegatronBertConfig](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertConfig) (MegatronBertConfig 모델)
- **mellum** -- [MellumConfig](/docs/transformers/v5.23.0/ko/model_doc/mellum#transformers.MellumConfig) (MellumConfig 모델)
- **metaclip_2** -- [MetaClip2Config](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2Config) (MetaClip2Config 모델)
- **metaclip_2_text_model** -- [MetaClip2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2TextConfig) (MetaClip2TextConfig 모델)
- **metaclip_2_vision_model** -- [MetaClip2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2VisionConfig) (MetaClip2VisionConfig 모델)
- **mgp-str** -- [MgpstrConfig](/docs/transformers/v5.23.0/ko/model_doc/mgp-str#transformers.MgpstrConfig) (MgpstrConfig 모델)
- **mimi** -- [MimiConfig](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiConfig) (MimiConfig 모델)
- **mimi_decoder** -- [MimiDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiDecoderConfig) (MimiDecoderConfig 모델)
- **mimi_encoder** -- [MimiEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiEncoderConfig) (MimiEncoderConfig 모델)
- **mimo_v2_flash** -- [MiMoV2FlashConfig](/docs/transformers/v5.23.0/ko/model_doc/mimo_v2_flash#transformers.MiMoV2FlashConfig) (MiMoV2FlashConfig 모델)
- **minicpm3** -- [MiniCPM3Config](/docs/transformers/v5.23.0/ko/model_doc/minicpm3#transformers.MiniCPM3Config) (MiniCPM3Config 모델)
- **minicpm_ocr** -- [MiniCPMOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrConfig) (MiniCPMOcrConfig 모델)
- **minicpm_ocr_vision** -- [MiniCPMOcrVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrVisionConfig) (MiniCPMOcrVisionConfig 모델)
- **minicpmv4_6** -- [MiniCPMV4_6Config](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6Config) (MiniCPMV4_6Config 모델)
- **minicpmv4_6_vision** -- [MiniCPMV4_6VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6VisionConfig) (MiniCPMV4_6VisionConfig 모델)
- **minicpmv4_7** -- [MiniCPMV4_7Config](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7Config) (MiniCPMV4_7Config 모델)
- **minicpmv4_7_vision** -- [MiniCPMV4_7VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7VisionConfig) (MiniCPMV4_7VisionConfig 모델)
- **minimax** -- [MiniMaxConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxConfig) (MiniMaxConfig 모델)
- **minimax_m2** -- [MiniMaxM2Config](/docs/transformers/v5.23.0/ko/model_doc/minimax_m2#transformers.MiniMaxM2Config) (MiniMaxM2Config 모델)
- **minimax_m3_vl** -- [MiniMaxM3VLConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLConfig) (MiniMaxM3VLConfig 모델)
- **minimax_m3_vl_text** -- [MiniMaxM3VLTextConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLTextConfig) (MiniMaxM3VLTextConfig 모델)
- **minimax_m3_vl_vision** -- [MiniMaxM3VLVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLVisionConfig) (MiniMaxM3VLVisionConfig 모델)
- **ministral** -- [MinistralConfig](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralConfig) (MinistralConfig 모델)
- **ministral3** -- [Ministral3Config](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3Config) (Ministral3Config 모델)
- **mistral** -- [MistralConfig](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralConfig) (MistralConfig 모델)
- **mistral3** -- [Mistral3Config](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3Config) (Mistral3Config 모델)
- **mistral4** -- [Mistral4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4Config) (Mistral4Config 모델)
- **mistral_large4** -- [MistralLarge4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4Config) (MistralLarge4Config 모델)
- **mistral_large4_text** -- [MistralLarge4TextConfig](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4TextConfig) (MistralLarge4TextConfig 모델)
- **mixtral** -- [MixtralConfig](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralConfig) (MixtralConfig 모델)
- **mlcd** -- [MLCDVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/mlcd#transformers.MLCDVisionConfig) (MLCDVisionConfig 모델)
- **mlcd_vision_model** -- [MLCDVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/mlcd#transformers.MLCDVisionConfig) (MLCDVisionConfig 모델)
- **mllama** -- [MllamaConfig](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaConfig) (MllamaConfig 모델)
- **mllama_text_model** -- [MllamaTextConfig](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaTextConfig) (MllamaTextConfig 모델)
- **mllama_vision_model** -- [MllamaVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaVisionConfig) (MllamaVisionConfig 모델)
- **mm-grounding-dino** -- [MMGroundingDinoConfig](/docs/transformers/v5.23.0/ko/model_doc/mm-grounding-dino#transformers.MMGroundingDinoConfig) (MMGroundingDinoConfig 모델)
- **mobilebert** -- [MobileBertConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertConfig) (MobileBertConfig 모델)
- **mobilenet_v1** -- [MobileNetV1Config](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v1#transformers.MobileNetV1Config) (MobileNetV1Config 모델)
- **mobilenet_v2** -- [MobileNetV2Config](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v2#transformers.MobileNetV2Config) (MobileNetV2Config 모델)
- **mobilevit** -- [MobileViTConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilevit#transformers.MobileViTConfig) (MobileViTConfig 모델)
- **mobilevitv2** -- [MobileViTV2Config](/docs/transformers/v5.23.0/ko/model_doc/mobilevitv2#transformers.MobileViTV2Config) (MobileViTV2Config 모델)
- **modernbert** -- [ModernBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertConfig) (ModernBertConfig 모델)
- **modernbert-decoder** -- [ModernBertDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/modernbert-decoder#transformers.ModernBertDecoderConfig) (ModernBertDecoderConfig 모델)
- **modernvbert** -- [ModernVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertConfig) (ModernVBertConfig 모델)
- **molmo2** -- [Molmo2Config](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2Config) (Molmo2Config 모델)
- **molmo2_adapter** -- [Molmo2AdapterConfig](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2AdapterConfig) (Molmo2AdapterConfig 모델)
- **molmo2_text** -- [Molmo2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2TextConfig) (Molmo2TextConfig 모델)
- **molmo2_vision** -- [Molmo2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2VisionConfig) (Molmo2VisionConfig 모델)
- **moonshine** -- [MoonshineConfig](/docs/transformers/v5.23.0/ko/model_doc/moonshine#transformers.MoonshineConfig) (MoonshineConfig 모델)
- **moonshine_streaming** -- [MoonshineStreamingConfig](/docs/transformers/v5.23.0/ko/model_doc/moonshine_streaming#transformers.MoonshineStreamingConfig) (MoonshineStreamingConfig 모델)
- **moonshine_streaming_encoder** -- [MoonshineStreamingEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/moonshine_streaming#transformers.MoonshineStreamingEncoderConfig) (MoonshineStreamingEncoderConfig 모델)
- **moshi** -- [MoshiConfig](/docs/transformers/v5.23.0/ko/model_doc/moshi#transformers.MoshiConfig) (MoshiConfig 모델)
- **moshi_depth** -- [MoshiDepthConfig](/docs/transformers/v5.23.0/ko/model_doc/moshi#transformers.MoshiDepthConfig) (MoshiDepthConfig 모델)
- **moss_transcribe_diarize** -- [MossTranscribeDiarizeConfig](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeConfig) (MossTranscribeDiarizeConfig 모델)
- **moss_transcribe_diarize_audio** -- [MossTranscribeDiarizeAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeAudioConfig) (MossTranscribeDiarizeAudioConfig 모델)
- **mpnet** -- [MPNetConfig](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetConfig) (MPNetConfig 모델)
- **mpt** -- [MptConfig](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptConfig) (MptConfig 모델)
- **mra** -- [MraConfig](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraConfig) (MraConfig 모델)
- **mt5** -- [MT5Config](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5Config) (MT5Config 모델)
- **muse_glimmer** -- [MuseGlimmerConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerConfig) (MuseGlimmerConfig 모델)
- **muse_glimmer_assistant** -- [MuseGlimmerAssistantConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer_assistant#transformers.MuseGlimmerAssistantConfig) (MuseGlimmerAssistantConfig 모델)
- **muse_glimmer_text** -- [MuseGlimmerTextConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerTextConfig) (MuseGlimmerTextConfig 모델)
- **muse_glimmer_vision** -- [MuseGlimmerVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerVisionConfig) (MuseGlimmerVisionConfig 모델)
- **muse_spark** -- [MuseSparkConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkConfig) (MuseSparkConfig 모델)
- **muse_spark_audio** -- [MuseSparkAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkAudioConfig) (MuseSparkAudioConfig 모델)
- **muse_spark_text** -- [MuseSparkTextConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkTextConfig) (MuseSparkTextConfig 모델)
- **muse_spark_vision** -- [MuseSparkVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkVisionConfig) (MuseSparkVisionConfig 모델)
- **musicflamingo** -- [MusicFlamingoConfig](/docs/transformers/v5.23.0/ko/model_doc/musicflamingo#transformers.MusicFlamingoConfig) (MusicFlamingoConfig 모델)
- **musicgen** -- [MusicgenConfig](/docs/transformers/v5.23.0/ko/model_doc/musicgen#transformers.MusicgenConfig) (MusicgenConfig 모델)
- **musicgen_decoder** -- [MusicgenDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/musicgen#transformers.MusicgenDecoderConfig) (MusicgenDecoderConfig 모델)
- **musicgen_melody** -- [MusicgenMelodyConfig](/docs/transformers/v5.23.0/ko/model_doc/musicgen_melody#transformers.MusicgenMelodyConfig) (MusicgenMelodyConfig 모델)
- **musicgen_melody_decoder** -- [MusicgenMelodyDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/musicgen_melody#transformers.MusicgenMelodyDecoderConfig) (MusicgenMelodyDecoderConfig 모델)
- **mvp** -- [MvpConfig](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpConfig) (MvpConfig 모델)
- **nanochat** -- [NanoChatConfig](/docs/transformers/v5.23.0/ko/model_doc/nanochat#transformers.NanoChatConfig) (NanoChatConfig 모델)
- **nemo_speechlm** -- [Nemotron3_5TranscribeConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeConfig) (Nemotron3_5TranscribeConfig 모델)
- **nemotron** -- [NemotronConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronConfig) (NemotronConfig 모델)
- **nemotron3_5_asr** -- [Nemotron3_5AsrConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_asr#transformers.Nemotron3_5AsrConfig) (Nemotron3_5AsrConfig 모델)
- **nemotron3_5_transcribe_asr_audio** -- [Nemotron3_5TranscribeAsrAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeAsrAudioConfig) (Nemotron3_5TranscribeAsrAudioConfig 모델)
- **nemotron3_5_transcribe_audio** -- [Nemotron3_5TranscribeAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeAudioConfig) (Nemotron3_5TranscribeAudioConfig 모델)
- **nemotron3_5_transcribe_word_timestamp** -- [Nemotron3_5TranscribeWordTimestampConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeWordTimestampConfig) (Nemotron3_5TranscribeWordTimestampConfig 모델)
- **nemotron3_diarization** -- [Nemotron3DiarizationConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationConfig) (Nemotron3DiarizationConfig 모델)
- **nemotron3_diarization_audio** -- [Nemotron3DiarizationAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationAudioConfig) (Nemotron3DiarizationAudioConfig 모델)
- **nemotron_asr_streaming** -- [NemotronAsrStreamingConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingConfig) (NemotronAsrStreamingConfig 모델)
- **nemotron_asr_streaming_encoder** -- [NemotronAsrStreamingEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingEncoderConfig) (NemotronAsrStreamingEncoderConfig 모델)
- **nemotron_h** -- [NemotronHConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h#transformers.NemotronHConfig) (NemotronHConfig 모델)
- **nemotron_h_omni** -- [NemotronH_Omni_Reasoning_V3_Config](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h_omni#transformers.NemotronH_Omni_Reasoning_V3_Config) (NemotronH_Omni_Reasoning_V3_Config 모델)
- **neomme** -- [NeoMMEConfig](/docs/transformers/v5.23.0/ko/model_doc/neomme#transformers.NeoMMEConfig) (NeoMMEConfig 모델)
- **neucodec** -- [NeuCodecConfig](/docs/transformers/v5.23.0/ko/model_doc/neucodec#transformers.NeuCodecConfig) (NeuCodecConfig 모델)
- **nllb-moe** -- [NllbMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/nllb-moe#transformers.NllbMoeConfig) (NllbMoeConfig 모델)
- **nomic_bert** -- [NomicBertConfig](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertConfig) (NomicBertConfig 모델)
- **nougat** -- [NougatConfig](/docs/transformers/v5.23.0/ko/model_doc/nougat#transformers.NougatConfig) (NougatConfig 모델)
- **nystromformer** -- [NystromformerConfig](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerConfig) (NystromformerConfig 모델)
- **olmo** -- [OlmoConfig](/docs/transformers/v5.23.0/ko/model_doc/olmo#transformers.OlmoConfig) (OlmoConfig 모델)
- **olmo2** -- [Olmo2Config](/docs/transformers/v5.23.0/ko/model_doc/olmo2#transformers.Olmo2Config) (Olmo2Config 모델)
- **olmo3** -- [Olmo3Config](/docs/transformers/v5.23.0/ko/model_doc/olmo3#transformers.Olmo3Config) (Olmo3Config 모델)
- **olmo_hybrid** -- [OlmoHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/olmo_hybrid#transformers.OlmoHybridConfig) (OlmoHybridConfig 모델)
- **olmoe** -- [OlmoeConfig](/docs/transformers/v5.23.0/ko/model_doc/olmoe#transformers.OlmoeConfig) (OlmoeConfig 모델)
- **omdet-turbo** -- [OmDetTurboConfig](/docs/transformers/v5.23.0/ko/model_doc/omdet-turbo#transformers.OmDetTurboConfig) (OmDetTurboConfig 모델)
- **oneformer** -- [OneFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/oneformer#transformers.OneFormerConfig) (OneFormerConfig 모델)
- **openai-gpt** -- [OpenAIGPTConfig](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTConfig) (OpenAIGPTConfig 모델)
- **openai_privacy_filter** -- [OpenAIPrivacyFilterConfig](/docs/transformers/v5.23.0/ko/model_doc/openai_privacy_filter#transformers.OpenAIPrivacyFilterConfig) (OpenAIPrivacyFilterConfig 모델)
- **opt** -- [OPTConfig](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTConfig) (OPTConfig 모델)
- **ovis2** -- [Ovis2Config](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2Config) (Ovis2Config 모델)
- **ovis2_vision** -- [Ovis2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2VisionConfig) (Ovis2VisionConfig 모델)
- **owlv2** -- [Owlv2Config](/docs/transformers/v5.23.0/ko/model_doc/owlv2#transformers.Owlv2Config) (Owlv2Config 모델)
- **owlv2_text_model** -- [Owlv2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/owlv2#transformers.Owlv2TextConfig) (Owlv2TextConfig 모델)
- **owlv2_vision_model** -- [Owlv2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/owlv2#transformers.Owlv2VisionConfig) (Owlv2VisionConfig 모델)
- **owlvit** -- [OwlViTConfig](/docs/transformers/v5.23.0/ko/model_doc/owlvit#transformers.OwlViTConfig) (OwlViTConfig 모델)
- **owlvit_text_model** -- [OwlViTTextConfig](/docs/transformers/v5.23.0/ko/model_doc/owlvit#transformers.OwlViTTextConfig) (OwlViTTextConfig 모델)
- **owlvit_vision_model** -- [OwlViTVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/owlvit#transformers.OwlViTVisionConfig) (OwlViTVisionConfig 모델)
- **paddleocr_vl** -- [PaddleOCRVLConfig](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVLConfig) (PaddleOCRVLConfig 모델)
- **paddleocr_vl_text** -- [PaddleOCRTextConfig](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRTextConfig) (PaddleOCRTextConfig 모델)
- **paddleocr_vl_vision** -- [PaddleOCRVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVisionConfig) (PaddleOCRVisionConfig 모델)
- **paligemma** -- [PaliGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaConfig) (PaliGemmaConfig 모델)
- **parakeet_ctc** -- [ParakeetCTCConfig](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetCTCConfig) (ParakeetCTCConfig 모델)
- **parakeet_encoder** -- [ParakeetEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetEncoderConfig) (ParakeetEncoderConfig 모델)
- **parakeet_rnnt** -- [ParakeetRNNTConfig](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetRNNTConfig) (ParakeetRNNTConfig 모델)
- **parakeet_tdt** -- [ParakeetTDTConfig](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetTDTConfig) (ParakeetTDTConfig 모델)
- **patchtsmixer** -- [PatchTSMixerConfig](/docs/transformers/v5.23.0/ko/model_doc/patchtsmixer#transformers.PatchTSMixerConfig) (PatchTSMixerConfig 모델)
- **patchtst** -- [PatchTSTConfig](/docs/transformers/v5.23.0/ko/model_doc/patchtst#transformers.PatchTSTConfig) (PatchTSTConfig 모델)
- **pe_audio** -- [PeAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_audio#transformers.PeAudioConfig) (PeAudioConfig 모델)
- **pe_audio_encoder** -- [PeAudioEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_audio#transformers.PeAudioEncoderConfig) (PeAudioEncoderConfig 모델)
- **pe_audio_video** -- [PeAudioVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_audio_video#transformers.PeAudioVideoConfig) (PeAudioVideoConfig 모델)
- **pe_audio_video_encoder** -- [PeAudioVideoEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_audio_video#transformers.PeAudioVideoEncoderConfig) (PeAudioVideoEncoderConfig 모델)
- **pe_video** -- [PeVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_video#transformers.PeVideoConfig) (PeVideoConfig 모델)
- **pe_video_encoder** -- [PeVideoEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_video#transformers.PeVideoEncoderConfig) (PeVideoEncoderConfig 모델)
- **pegasus** -- [PegasusConfig](/docs/transformers/v5.23.0/ko/model_doc/pegasus#transformers.PegasusConfig) (PegasusConfig 모델)
- **pegasus_x** -- [PegasusXConfig](/docs/transformers/v5.23.0/ko/model_doc/pegasus_x#transformers.PegasusXConfig) (PegasusXConfig 모델)
- **perceiver** -- [PerceiverConfig](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverConfig) (PerceiverConfig 모델)
- **perception_lm** -- [PerceptionLMConfig](/docs/transformers/v5.23.0/ko/model_doc/perception_lm#transformers.PerceptionLMConfig) (PerceptionLMConfig 모델)
- **persimmon** -- [PersimmonConfig](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonConfig) (PersimmonConfig 모델)
- **phi** -- [PhiConfig](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiConfig) (PhiConfig 모델)
- **phi3** -- [Phi3Config](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3Config) (Phi3Config 모델)
- **phi4_multimodal** -- [Phi4MultimodalConfig](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalConfig) (Phi4MultimodalConfig 모델)
- **phi4_multimodal_audio** -- [Phi4MultimodalAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalAudioConfig) (Phi4MultimodalAudioConfig 모델)
- **phi4_multimodal_vision** -- [Phi4MultimodalVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalVisionConfig) (Phi4MultimodalVisionConfig 모델)
- **phimoe** -- [PhimoeConfig](/docs/transformers/v5.23.0/ko/model_doc/phimoe#transformers.PhimoeConfig) (PhimoeConfig 모델)
- **pi0** -- [PI0Config](/docs/transformers/v5.23.0/ko/model_doc/pi0#transformers.PI0Config) (PI0Config 모델)
- **pix2struct** -- [Pix2StructConfig](/docs/transformers/v5.23.0/ko/model_doc/pix2struct#transformers.Pix2StructConfig) (Pix2StructConfig 모델)
- **pix2struct_text_model** -- [Pix2StructTextConfig](/docs/transformers/v5.23.0/ko/model_doc/pix2struct#transformers.Pix2StructTextConfig) (Pix2StructTextConfig 모델)
- **pix2struct_vision_model** -- [Pix2StructVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/pix2struct#transformers.Pix2StructVisionConfig) (Pix2StructVisionConfig 모델)
- **pixio** -- [PixioConfig](/docs/transformers/v5.23.0/ko/model_doc/pixio#transformers.PixioConfig) (PixioConfig 모델)
- **pixtral** -- [PixtralVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/pixtral#transformers.PixtralVisionConfig) (PixtralVisionConfig 모델)
- **plbart** -- [PLBartConfig](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartConfig) (PLBartConfig 모델)
- **pocket_tts** -- [PocketTTSConfig](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts#transformers.PocketTTSConfig) (PocketTTSConfig 모델)
- **pocket_tts_audio_tokenizer** -- [PocketTTSAudioTokenizerConfig](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerConfig) (PocketTTSAudioTokenizerConfig 모델)
- **pocket_tts_audio_tokenizer_decoder** -- [PocketTTSAudioTokenizerDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerDecoderConfig) (PocketTTSAudioTokenizerDecoderConfig 모델)
- **pocket_tts_audio_tokenizer_encoder** -- [PocketTTSAudioTokenizerEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerEncoderConfig) (PocketTTSAudioTokenizerEncoderConfig 모델)
- **poolformer** -- [PoolFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/poolformer#transformers.PoolFormerConfig) (PoolFormerConfig 모델)
- **pop2piano** -- [Pop2PianoConfig](/docs/transformers/v5.23.0/ko/model_doc/pop2piano#transformers.Pop2PianoConfig) (Pop2PianoConfig 모델)
- **pp_chart2table** -- [PPChart2TableConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_chart2table#transformers.PPChart2TableConfig) (PPChart2TableConfig 모델)
- **pp_chart2table_vision** -- [PPChart2TableVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_chart2table#transformers.PPChart2TableVisionConfig) (PPChart2TableVisionConfig 모델)
- **pp_doclayout_v2** -- [PPDocLayoutV2Config](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v2#transformers.PPDocLayoutV2Config) (PPDocLayoutV2Config 모델)
- **pp_doclayout_v3** -- [PPDocLayoutV3Config](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v3#transformers.PPDocLayoutV3Config) (PPDocLayoutV3Config 모델)
- **pp_doclayout_v4** -- [PPDocLayoutV4Config](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v4#transformers.PPDocLayoutV4Config) (PPDocLayoutV4Config 모델)
- **pp_formulanet** -- [PPFormulaNetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_formulanet#transformers.PPFormulaNetConfig) (PPFormulaNetConfig 모델)
- **pp_formulanet_text** -- [PPFormulaNetTextConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_formulanet#transformers.PPFormulaNetTextConfig) (PPFormulaNetTextConfig 모델)
- **pp_formulanet_vision** -- [PPFormulaNetVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_formulanet#transformers.PPFormulaNetVisionConfig) (PPFormulaNetVisionConfig 모델)
- **pp_lcnet** -- [PPLCNetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_lcnet#transformers.PPLCNetConfig) (PPLCNetConfig 모델)
- **pp_lcnet_v3** -- [PPLCNetV3Config](/docs/transformers/v5.23.0/ko/model_doc/pp_lcnet_v3#transformers.PPLCNetV3Config) (PPLCNetV3Config 모델)
- **pp_lcnet_v4** -- [PPLCNetV4Config](/docs/transformers/v5.23.0/ko/model_doc/pp_lcnet_v4#transformers.PPLCNetV4Config) (PPLCNetV4Config 모델)
- **pp_ocrv5_mobile_det** -- [PPOCRV5MobileDetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_mobile_det#transformers.PPOCRV5MobileDetConfig) (PPOCRV5MobileDetConfig 모델)
- **pp_ocrv5_mobile_rec** -- [PPOCRV5MobileRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_mobile_rec#transformers.PPOCRV5MobileRecConfig) (PPOCRV5MobileRecConfig 모델)
- **pp_ocrv5_server_det** -- [PPOCRV5ServerDetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_server_det#transformers.PPOCRV5ServerDetConfig) (PPOCRV5ServerDetConfig 모델)
- **pp_ocrv5_server_rec** -- [PPOCRV5ServerRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_server_rec#transformers.PPOCRV5ServerRecConfig) (PPOCRV5ServerRecConfig 모델)
- **pp_ocrv6_medium_det** -- [PPOCRV6MediumDetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_medium_det#transformers.PPOCRV6MediumDetConfig) (PPOCRV6MediumDetConfig 모델)
- **pp_ocrv6_small_det** -- [PPOCRV6SmallDetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_det#transformers.PPOCRV6SmallDetConfig) (PPOCRV6SmallDetConfig 모델)
- **pp_ocrv6_small_rec** -- [PPOCRV6SmallRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_rec#transformers.PPOCRV6SmallRecConfig) (PPOCRV6SmallRecConfig 모델)
- **pp_ocrv6_tiny_rec** -- [PPOCRV6TinyRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_tiny_rec#transformers.PPOCRV6TinyRecConfig) (PPOCRV6TinyRecConfig 모델)
- **prompt_depth_anything** -- [PromptDepthAnythingConfig](/docs/transformers/v5.23.0/ko/model_doc/prompt_depth_anything#transformers.PromptDepthAnythingConfig) (PromptDepthAnythingConfig 모델)
- **prophetnet** -- [ProphetNetConfig](/docs/transformers/v5.23.0/ko/model_doc/prophetnet#transformers.ProphetNetConfig) (ProphetNetConfig 모델)
- **pvt** -- [PvtConfig](/docs/transformers/v5.23.0/ko/model_doc/pvt#transformers.PvtConfig) (PvtConfig 모델)
- **pvt_v2** -- [PvtV2Config](/docs/transformers/v5.23.0/ko/model_doc/pvt_v2#transformers.PvtV2Config) (PvtV2Config 모델)
- **qianfan_ocr** -- [QianfanOCRConfig](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRConfig) (QianfanOCRConfig 모델)
- **qianfan_ocr_vision** -- [QianfanOCRVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRVisionConfig) (QianfanOCRVisionConfig 모델)
- **qwen2** -- [Qwen2Config](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Config) (Qwen2Config 모델)
- **qwen2_5_omni** -- [Qwen2_5OmniConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniConfig) (Qwen2_5OmniConfig 모델)
- **qwen2_5_omni_audio_encoder** -- [Qwen2_5OmniAudioEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniAudioEncoderConfig) (Qwen2_5OmniAudioEncoderConfig 모델)
- **qwen2_5_omni_bigvgan** -- [Qwen2_5OmniBigVGANConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniBigVGANConfig) (Qwen2_5OmniBigVGANConfig 모델)
- **qwen2_5_omni_dit** -- [Qwen2_5OmniDiTConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniDiTConfig) (Qwen2_5OmniDiTConfig 모델)
- **qwen2_5_omni_talker** -- [Qwen2_5OmniTalkerConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniTalkerConfig) (Qwen2_5OmniTalkerConfig 모델)
- **qwen2_5_omni_text** -- [Qwen2_5OmniTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniTextConfig) (Qwen2_5OmniTextConfig 모델)
- **qwen2_5_omni_thinker** -- [Qwen2_5OmniThinkerConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniThinkerConfig) (Qwen2_5OmniThinkerConfig 모델)
- **qwen2_5_omni_token2wav** -- [Qwen2_5OmniToken2WavConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniToken2WavConfig) (Qwen2_5OmniToken2WavConfig 모델)
- **qwen2_5_omni_vision_encoder** -- [Qwen2_5OmniVisionEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniVisionEncoderConfig) (Qwen2_5OmniVisionEncoderConfig 모델)
- **qwen2_5_vl** -- [Qwen2_5_VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLConfig) (Qwen2_5_VLConfig 모델)
- **qwen2_5_vl_text** -- [Qwen2_5_VLTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLTextConfig) (Qwen2_5_VLTextConfig 모델)
- **qwen2_5_vl_vision** -- [Qwen2_5_VLVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLVisionConfig) (Qwen2_5_VLVisionConfig 모델)
- **qwen2_audio** -- [Qwen2AudioConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioConfig) (Qwen2AudioConfig 모델)
- **qwen2_audio_encoder** -- [Qwen2AudioEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioEncoderConfig) (Qwen2AudioEncoderConfig 모델)
- **qwen2_moe** -- [Qwen2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeConfig) (Qwen2MoeConfig 모델)
- **qwen2_vl** -- [Qwen2VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLConfig) (Qwen2VLConfig 모델)
- **qwen2_vl_text** -- [Qwen2VLTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLTextConfig) (Qwen2VLTextConfig 모델)
- **qwen2_vl_vision** -- [Qwen2VLVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLVisionConfig) (Qwen2VLVisionConfig 모델)
- **qwen3** -- [Qwen3Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3Config) (Qwen3Config 모델)
- **qwen3_5** -- [Qwen3_5Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Config) (Qwen3_5Config 모델)
- **qwen3_5_moe** -- [Qwen3_5MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeConfig) (Qwen3_5MoeConfig 모델)
- **qwen3_5_moe_text** -- [Qwen3_5MoeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeTextConfig) (Qwen3_5MoeTextConfig 모델)
- **qwen3_5_moe_vision** -- [Qwen3_5MoeVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeVisionConfig) (Qwen3_5MoeVisionConfig 모델)
- **qwen3_5_text** -- [Qwen3_5TextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5TextConfig) (Qwen3_5TextConfig 모델)
- **qwen3_5_vision** -- [Qwen3_5VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5VisionConfig) (Qwen3_5VisionConfig 모델)
- **qwen3_asr** -- [Qwen3ASRConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRConfig) (Qwen3ASRConfig 모델)
- **qwen3_asr_encoder** -- [Qwen3ASREncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASREncoderConfig) (Qwen3ASREncoderConfig 모델)
- **qwen3_moe** -- [Qwen3MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeConfig) (Qwen3MoeConfig 모델)
- **qwen3_next** -- [Qwen3NextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextConfig) (Qwen3NextConfig 모델)
- **qwen3_omni_moe** -- [Qwen3OmniMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeConfig) (Qwen3OmniMoeConfig 모델)
- **qwen3_omni_moe_audio_encoder** -- [Qwen3OmniMoeAudioEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeAudioEncoderConfig) (Qwen3OmniMoeAudioEncoderConfig 모델)
- **qwen3_omni_moe_talker_code_predictor** -- [Qwen3OmniMoeTalkerCodePredictorConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeTalkerCodePredictorConfig) (Qwen3OmniMoeTalkerCodePredictorConfig 모델)
- **qwen3_omni_moe_talker_text** -- [Qwen3OmniMoeTalkerTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeTalkerTextConfig) (Qwen3OmniMoeTalkerTextConfig 모델)
- **qwen3_omni_moe_text** -- [Qwen3OmniMoeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeTextConfig) (Qwen3OmniMoeTextConfig 모델)
- **qwen3_omni_moe_thinker** -- [Qwen3OmniMoeThinkerConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeThinkerConfig) (Qwen3OmniMoeThinkerConfig 모델)
- **qwen3_omni_moe_vision_encoder** -- [Qwen3OmniMoeVisionEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeVisionEncoderConfig) (Qwen3OmniMoeVisionEncoderConfig 모델)
- **qwen3_vl** -- [Qwen3VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLConfig) (Qwen3VLConfig 모델)
- **qwen3_vl_moe** -- [Qwen3VLMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeConfig) (Qwen3VLMoeConfig 모델)
- **qwen3_vl_moe_text** -- [Qwen3VLMoeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeTextConfig) (Qwen3VLMoeTextConfig 모델)
- **qwen3_vl_moe_vision** -- [Qwen3VLMoeVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeVisionConfig) (Qwen3VLMoeVisionConfig 모델)
- **qwen3_vl_text** -- [Qwen3VLTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLTextConfig) (Qwen3VLTextConfig 모델)
- **qwen3_vl_vision** -- [Qwen3VLVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLVisionConfig) (Qwen3VLVisionConfig 모델)
- **qwen4_exp** -- [Qwen4ExpConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpConfig) (Qwen4ExpConfig 모델)
- **qwen4_exp_text** -- [Qwen4ExpTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpTextConfig) (Qwen4ExpTextConfig 모델)
- **qwen4_exp_vision** -- [Qwen4ExpVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpVisionConfig) (Qwen4ExpVisionConfig 모델)
- **radio** -- [RadioConfig](/docs/transformers/v5.23.0/ko/model_doc/radio#transformers.RadioConfig) (RadioConfig 모델)
- **rag** -- [RagConfig](/docs/transformers/v5.23.0/ko/model_doc/rag#transformers.RagConfig) (RagConfig 모델)
- **recurrent_gemma** -- [RecurrentGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/recurrent_gemma#transformers.RecurrentGemmaConfig) (RecurrentGemmaConfig 모델)
- **reformer** -- [ReformerConfig](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerConfig) (ReformerConfig 모델)
- **regnet** -- [RegNetConfig](/docs/transformers/v5.23.0/ko/model_doc/regnet#transformers.RegNetConfig) (RegNetConfig 모델)
- **rembert** -- [RemBertConfig](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertConfig) (RemBertConfig 모델)
- **resnet** -- [ResNetConfig](/docs/transformers/v5.23.0/ko/model_doc/resnet#transformers.ResNetConfig) (ResNetConfig 모델)
- **rf_detr** -- [RfDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/rf_detr#transformers.RfDetrConfig) (RfDetrConfig 모델)
- **rf_detr_dinov2** -- [RfDetrDinov2Config](/docs/transformers/v5.23.0/ko/model_doc/rf_detr#transformers.RfDetrDinov2Config) (RfDetrDinov2Config 모델)
- **roberta** -- [RobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaConfig) (RobertaConfig 모델)
- **roberta-prelayernorm** -- [RobertaPreLayerNormConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormConfig) (RobertaPreLayerNormConfig 모델)
- **roc_bert** -- [RoCBertConfig](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertConfig) (RoCBertConfig 모델)
- **roformer** -- [RoFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerConfig) (RoFormerConfig 모델)
- **rt_detr** -- [RTDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/rt_detr#transformers.RTDetrConfig) (RTDetrConfig 모델)
- **rt_detr_resnet** -- [RTDetrResNetConfig](/docs/transformers/v5.23.0/ko/model_doc/rt_detr#transformers.RTDetrResNetConfig) (RTDetrResNetConfig 모델)
- **rt_detr_v2** -- [RTDetrV2Config](/docs/transformers/v5.23.0/ko/model_doc/rt_detr_v2#transformers.RTDetrV2Config) (RTDetrV2Config 모델)
- **rwkv** -- [RwkvConfig](/docs/transformers/v5.23.0/ko/model_doc/rwkv#transformers.RwkvConfig) (RwkvConfig 모델)
- **sam** -- [SamConfig](/docs/transformers/v5.23.0/ko/model_doc/sam#transformers.SamConfig) (SamConfig 모델)
- **sam2** -- [Sam2Config](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2Config) (Sam2Config 모델)
- **sam2_hiera_det_model** -- [Sam2HieraDetConfig](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2HieraDetConfig) (Sam2HieraDetConfig 모델)
- **sam2_mask_decoder** -- [Sam2MaskDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2MaskDecoderConfig) (Sam2MaskDecoderConfig 모델)
- **sam2_prompt_encoder** -- [Sam2PromptEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2PromptEncoderConfig) (Sam2PromptEncoderConfig 모델)
- **sam2_video** -- [Sam2VideoConfig](/docs/transformers/v5.23.0/ko/model_doc/sam2_video#transformers.Sam2VideoConfig) (Sam2VideoConfig 모델)
- **sam2_video_mask_decoder** -- [Sam2VideoMaskDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam2_video#transformers.Sam2VideoMaskDecoderConfig) (Sam2VideoMaskDecoderConfig 모델)
- **sam2_video_prompt_encoder** -- [Sam2VideoPromptEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam2_video#transformers.Sam2VideoPromptEncoderConfig) (Sam2VideoPromptEncoderConfig 모델)
- **sam2_vision_model** -- [Sam2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2VisionConfig) (Sam2VisionConfig 모델)
- **sam3** -- [Sam3Config](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3Config) (Sam3Config 모델)
- **sam3_detr_decoder** -- [Sam3DETRDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3DETRDecoderConfig) (Sam3DETRDecoderConfig 모델)
- **sam3_detr_encoder** -- [Sam3DETREncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3DETREncoderConfig) (Sam3DETREncoderConfig 모델)
- **sam3_geometry_encoder** -- [Sam3GeometryEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3GeometryEncoderConfig) (Sam3GeometryEncoderConfig 모델)
- **sam3_lite_text** -- [Sam3LiteTextConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextConfig) (Sam3LiteTextConfig 모델)
- **sam3_lite_text_detr_decoder** -- [Sam3LiteTextDETRDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextDETRDecoderConfig) (Sam3LiteTextDETRDecoderConfig 모델)
- **sam3_lite_text_detr_encoder** -- [Sam3LiteTextDETREncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextDETREncoderConfig) (Sam3LiteTextDETREncoderConfig 모델)
- **sam3_lite_text_geometry_encoder** -- [Sam3LiteTextGeometryEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextGeometryEncoderConfig) (Sam3LiteTextGeometryEncoderConfig 모델)
- **sam3_lite_text_mask_decoder** -- [Sam3LiteTextMaskDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextMaskDecoderConfig) (Sam3LiteTextMaskDecoderConfig 모델)
- **sam3_lite_text_text_model** -- [Sam3LiteTextTextConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextTextConfig) (Sam3LiteTextTextConfig 모델)
- **sam3_mask_decoder** -- [Sam3MaskDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3MaskDecoderConfig) (Sam3MaskDecoderConfig 모델)
- **sam3_tracker** -- [Sam3TrackerConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker#transformers.Sam3TrackerConfig) (Sam3TrackerConfig 모델)
- **sam3_tracker_mask_decoder** -- [Sam3TrackerMaskDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker#transformers.Sam3TrackerMaskDecoderConfig) (Sam3TrackerMaskDecoderConfig 모델)
- **sam3_tracker_prompt_encoder** -- [Sam3TrackerPromptEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker#transformers.Sam3TrackerPromptEncoderConfig) (Sam3TrackerPromptEncoderConfig 모델)
- **sam3_tracker_video** -- [Sam3TrackerVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker_video#transformers.Sam3TrackerVideoConfig) (Sam3TrackerVideoConfig 모델)
- **sam3_tracker_video_mask_decoder** -- [Sam3TrackerVideoMaskDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker_video#transformers.Sam3TrackerVideoMaskDecoderConfig) (Sam3TrackerVideoMaskDecoderConfig 모델)
- **sam3_tracker_video_prompt_encoder** -- [Sam3TrackerVideoPromptEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker_video#transformers.Sam3TrackerVideoPromptEncoderConfig) (Sam3TrackerVideoPromptEncoderConfig 모델)
- **sam3_video** -- [Sam3VideoConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_video#transformers.Sam3VideoConfig) (Sam3VideoConfig 모델)
- **sam3_vision_model** -- [Sam3VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3VisionConfig) (Sam3VisionConfig 모델)
- **sam3_vit_model** -- [Sam3ViTConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3ViTConfig) (Sam3ViTConfig 모델)
- **sam_hq** -- [SamHQConfig](/docs/transformers/v5.23.0/ko/model_doc/sam_hq#transformers.SamHQConfig) (SamHQConfig 모델)
- **sam_hq_mask_decoder** -- [SamHQMaskDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam_hq#transformers.SamHQMaskDecoderConfig) (SamHQMaskDecoderConfig 모델)
- **sam_hq_prompt_encoder** -- [SamHQPromptEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam_hq#transformers.SamHQPromptEncoderConfig) (SamHQPromptEncoderConfig 모델)
- **sam_hq_vision_model** -- [SamHQVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/sam_hq#transformers.SamHQVisionConfig) (SamHQVisionConfig 모델)
- **sam_mask_decoder** -- [SamMaskDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam#transformers.SamMaskDecoderConfig) (SamMaskDecoderConfig 모델)
- **sam_prompt_encoder** -- [SamPromptEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/sam#transformers.SamPromptEncoderConfig) (SamPromptEncoderConfig 모델)
- **sam_vision_model** -- [SamVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/sam#transformers.SamVisionConfig) (SamVisionConfig 모델)
- **sapiens2** -- [Sapiens2Config](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2Config) (Sapiens2Config 모델)
- **sapiens2_head** -- [Sapiens2HeadConfig](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2HeadConfig) (Sapiens2HeadConfig 모델)
- **seamless_m4t** -- [SeamlessM4TConfig](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TConfig) (SeamlessM4TConfig 모델)
- **seamless_m4t_v2** -- [SeamlessM4Tv2Config](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t_v2#transformers.SeamlessM4Tv2Config) (SeamlessM4Tv2Config 모델)
- **seed_oss** -- [SeedOssConfig](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssConfig) (SeedOssConfig 모델)
- **segformer** -- [SegformerConfig](/docs/transformers/v5.23.0/ko/model_doc/segformer#transformers.SegformerConfig) (SegformerConfig 모델)
- **seggpt** -- [SegGptConfig](/docs/transformers/v5.23.0/ko/model_doc/seggpt#transformers.SegGptConfig) (SegGptConfig 모델)
- **sew** -- [SEWConfig](/docs/transformers/v5.23.0/ko/model_doc/sew#transformers.SEWConfig) (SEWConfig 모델)
- **sew-d** -- [SEWDConfig](/docs/transformers/v5.23.0/ko/model_doc/sew-d#transformers.SEWDConfig) (SEWDConfig 모델)
- **shieldgemma2** -- [ShieldGemma2Config](/docs/transformers/v5.23.0/ko/model_doc/shieldgemma2#transformers.ShieldGemma2Config) (ShieldGemma2Config 모델)
- **siglip** -- [SiglipConfig](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipConfig) (SiglipConfig 모델)
- **siglip2** -- [Siglip2Config](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2Config) (Siglip2Config 모델)
- **siglip2_text_model** -- [Siglip2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2TextConfig) (Siglip2TextConfig 모델)
- **siglip2_vision_model** -- [Siglip2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2VisionConfig) (Siglip2VisionConfig 모델)
- **siglip_text_model** -- [SiglipTextConfig](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipTextConfig) (SiglipTextConfig 모델)
- **siglip_vision_model** -- [SiglipVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipVisionConfig) (SiglipVisionConfig 모델)
- **slanet** -- [SLANetConfig](/docs/transformers/v5.23.0/ko/model_doc/slanet#transformers.SLANetConfig) (SLANetConfig 모델)
- **slanext** -- [SLANeXtConfig](/docs/transformers/v5.23.0/ko/model_doc/slanext#transformers.SLANeXtConfig) (SLANeXtConfig 모델)
- **slanext_vision** -- [SLANeXtVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/slanext#transformers.SLANeXtVisionConfig) (SLANeXtVisionConfig 모델)
- **smollm3** -- [SmolLM3Config](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3Config) (SmolLM3Config 모델)
- **smolvlm** -- [SmolVLMConfig](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMConfig) (SmolVLMConfig 모델)
- **smolvlm_vision** -- [SmolVLMVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMVisionConfig) (SmolVLMVisionConfig 모델)
- **solar_open** -- [SolarOpenConfig](/docs/transformers/v5.23.0/ko/model_doc/solar_open#transformers.SolarOpenConfig) (SolarOpenConfig 모델)
- **speech-encoder-decoder** -- [SpeechEncoderDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/speech-encoder-decoder#transformers.SpeechEncoderDecoderConfig) (SpeechEncoderDecoderConfig 모델)
- **speech_to_text** -- [Speech2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/speech_to_text#transformers.Speech2TextConfig) (Speech2TextConfig 모델)
- **speecht5** -- [SpeechT5Config](/docs/transformers/v5.23.0/ko/model_doc/speecht5#transformers.SpeechT5Config) (SpeechT5Config 모델)
- **speecht5_hifigan** -- [SpeechT5HifiGanConfig](/docs/transformers/v5.23.0/ko/model_doc/speecht5#transformers.SpeechT5HifiGanConfig) (SpeechT5HifiGanConfig 모델)
- **splinter** -- [SplinterConfig](/docs/transformers/v5.23.0/ko/model_doc/splinter#transformers.SplinterConfig) (SplinterConfig 모델)
- **squeezebert** -- [SqueezeBertConfig](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertConfig) (SqueezeBertConfig 모델)
- **stablelm** -- [StableLmConfig](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmConfig) (StableLmConfig 모델)
- **starcoder2** -- [Starcoder2Config](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2Config) (Starcoder2Config 모델)
- **step3p5** -- [Step3p7TextConfig](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7TextConfig) (Step3p7TextConfig 모델)
- **step3p5_vision** -- [Step3p7VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7VisionConfig) (Step3p7VisionConfig 모델)
- **step3p7** -- [Step3p7Config](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7Config) (Step3p7Config 모델)
- **superglue** -- [SuperGlueConfig](/docs/transformers/v5.23.0/ko/model_doc/superglue#transformers.SuperGlueConfig) (SuperGlueConfig 모델)
- **superpoint** -- [SuperPointConfig](/docs/transformers/v5.23.0/ko/model_doc/superpoint#transformers.SuperPointConfig) (SuperPointConfig 모델)
- **swiftformer** -- [SwiftFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/swiftformer#transformers.SwiftFormerConfig) (SwiftFormerConfig 모델)
- **swin** -- [SwinConfig](/docs/transformers/v5.23.0/ko/model_doc/swin#transformers.SwinConfig) (SwinConfig 모델)
- **swin2sr** -- [Swin2SRConfig](/docs/transformers/v5.23.0/ko/model_doc/swin2sr#transformers.Swin2SRConfig) (Swin2SRConfig 모델)
- **swinv2** -- [Swinv2Config](/docs/transformers/v5.23.0/ko/model_doc/swinv2#transformers.Swinv2Config) (Swinv2Config 모델)
- **switch_transformers** -- [SwitchTransformersConfig](/docs/transformers/v5.23.0/ko/model_doc/switch_transformers#transformers.SwitchTransformersConfig) (SwitchTransformersConfig 모델)
- **t5** -- [T5Config](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Config) (T5Config 모델)
- **t5_gemma_module** -- [T5GemmaModuleConfig](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaModuleConfig) (T5GemmaModuleConfig 모델)
- **t5gemma** -- [T5GemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaConfig) (T5GemmaConfig 모델)
- **t5gemma2** -- [T5Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2Config) (T5Gemma2Config 모델)
- **t5gemma2_decoder** -- [T5Gemma2DecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2DecoderConfig) (T5Gemma2DecoderConfig 모델)
- **t5gemma2_encoder** -- [T5Gemma2EncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2EncoderConfig) (T5Gemma2EncoderConfig 모델)
- **t5gemma2_text** -- [T5Gemma2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2TextConfig) (T5Gemma2TextConfig 모델)
- **table-transformer** -- [TableTransformerConfig](/docs/transformers/v5.23.0/ko/model_doc/table-transformer#transformers.TableTransformerConfig) (TableTransformerConfig 모델)
- **tapas** -- [TapasConfig](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasConfig) (TapasConfig 모델)
- **textnet** -- [TextNetConfig](/docs/transformers/v5.23.0/ko/model_doc/textnet#transformers.TextNetConfig) (TextNetConfig 모델)
- **time_series_transformer** -- [TimeSeriesTransformerConfig](/docs/transformers/v5.23.0/ko/model_doc/time_series_transformer#transformers.TimeSeriesTransformerConfig) (TimeSeriesTransformerConfig 모델)
- **timesfm** -- [TimesFmConfig](/docs/transformers/v5.23.0/ko/model_doc/timesfm#transformers.TimesFmConfig) (TimesFmConfig 모델)
- **timesfm2_5** -- [TimesFm2_5Config](/docs/transformers/v5.23.0/ko/model_doc/timesfm2_5#transformers.TimesFm2_5Config) (TimesFm2_5Config 모델)
- **timesformer** -- [TimesformerConfig](/docs/transformers/v5.23.0/ko/model_doc/timesformer#transformers.TimesformerConfig) (TimesformerConfig 모델)
- **timm_backbone** -- [TimmBackboneConfig](/docs/transformers/v5.23.0/ko/main_classes/backbones#transformers.TimmBackboneConfig) (TimmBackboneConfig 모델)
- **timm_wrapper** -- [TimmWrapperConfig](/docs/transformers/v5.23.0/ko/model_doc/timm_wrapper#transformers.TimmWrapperConfig) (TimmWrapperConfig 모델)
- **tipsv2** -- [Tipsv2Config](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2Config) (Tipsv2Config 모델)
- **tipsv2_dpt** -- [Tipsv2DptConfig](/docs/transformers/v5.23.0/ko/model_doc/tipsv2_dpt#transformers.Tipsv2DptConfig) (Tipsv2DptConfig 모델)
- **tipsv2_text_model** -- [Tipsv2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2TextConfig) (Tipsv2TextConfig 모델)
- **tipsv2_vision_model** -- [Tipsv2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2VisionConfig) (Tipsv2VisionConfig 모델)
- **trocr** -- [TrOCRConfig](/docs/transformers/v5.23.0/ko/model_doc/trocr#transformers.TrOCRConfig) (TrOCRConfig 모델)
- **tvp** -- [TvpConfig](/docs/transformers/v5.23.0/ko/model_doc/tvp#transformers.TvpConfig) (TvpConfig 모델)
- **udop** -- [UdopConfig](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopConfig) (UdopConfig 모델)
- **ultrabert** -- [UltrabertConfig](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertConfig) (UltrabertConfig 모델)
- **ultravbert** -- [UltraVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertConfig) (UltraVBertConfig 모델)
- **umt5** -- [UMT5Config](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5Config) (UMT5Config 모델)
- **unispeech** -- [UniSpeechConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechConfig) (UniSpeechConfig 모델)
- **unispeech-sat** -- [UniSpeechSatConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatConfig) (UniSpeechSatConfig 모델)
- **univnet** -- [UnivNetConfig](/docs/transformers/v5.23.0/ko/model_doc/univnet#transformers.UnivNetConfig) (UnivNetConfig 모델)
- **unlimited_ocr** -- [UnlimitedOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrConfig) (UnlimitedOcrConfig 모델)
- **unlimited_ocr_sam_vision_model** -- [UnlimitedOcrSamVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrSamVisionConfig) (UnlimitedOcrSamVisionConfig 모델)
- **unlimited_ocr_text** -- [UnlimitedOcrTextConfig](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrTextConfig) (UnlimitedOcrTextConfig 모델)
- **unlimited_ocr_vision** -- [UnlimitedOcrVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrVisionConfig) (UnlimitedOcrVisionConfig 모델)
- **unlimited_ocr_vision_encoder** -- [UnlimitedOcrVisionEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrVisionEncoderConfig) (UnlimitedOcrVisionEncoderConfig 모델)
- **upernet** -- [UperNetConfig](/docs/transformers/v5.23.0/ko/model_doc/upernet#transformers.UperNetConfig) (UperNetConfig 모델)
- **uvdoc** -- [UVDocConfig](/docs/transformers/v5.23.0/ko/model_doc/uvdoc#transformers.UVDocConfig) (UVDocConfig 모델)
- **uvdoc_backbone** -- [UVDocBackboneConfig](/docs/transformers/v5.23.0/ko/model_doc/uvdoc#transformers.UVDocBackboneConfig) (UVDocBackboneConfig 모델)
- **vaultgemma** -- [VaultGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/vaultgemma#transformers.VaultGemmaConfig) (VaultGemmaConfig 모델)
- **vibevoice** -- [VibeVoiceConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice#transformers.VibeVoiceConfig) (VibeVoiceConfig 모델)
- **vibevoice_acoustic_tokenizer** -- [VibeVoiceAcousticTokenizerConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerConfig) (VibeVoiceAcousticTokenizerConfig 모델)
- **vibevoice_acoustic_tokenizer_decoder** -- [VibeVoiceAcousticTokenizerDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerDecoderConfig) (VibeVoiceAcousticTokenizerDecoderConfig 모델)
- **vibevoice_acoustic_tokenizer_encoder** -- [VibeVoiceAcousticTokenizerEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerEncoderConfig) (VibeVoiceAcousticTokenizerEncoderConfig 모델)
- **vibevoice_asr** -- [VibeVoiceAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrConfig) (VibeVoiceAsrConfig 모델)
- **vibevoice_diffusion** -- [VibeVoiceDiffusionHeadConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice#transformers.VibeVoiceDiffusionHeadConfig) (VibeVoiceDiffusionHeadConfig 모델)
- **video_llama_3** -- [VideoLlama3Config](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3Config) (VideoLlama3Config 모델)
- **video_llama_3_vision** -- [VideoLlama3VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3VisionConfig) (VideoLlama3VisionConfig 모델)
- **video_llava** -- [VideoLlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaConfig) (VideoLlavaConfig 모델)
- **videomae** -- [VideoMAEConfig](/docs/transformers/v5.23.0/ko/model_doc/videomae#transformers.VideoMAEConfig) (VideoMAEConfig 모델)
- **videomt** -- [VideomtConfig](/docs/transformers/v5.23.0/ko/model_doc/videomt#transformers.VideomtConfig) (VideomtConfig 모델)
- **videoprism** -- [VideoPrismConfig](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismConfig) (VideoPrismConfig 모델)
- **videoprism_text_model** -- [VideoPrismTextConfig](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismTextConfig) (VideoPrismTextConfig 모델)
- **videoprism_vision_model** -- [VideoPrismVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismVisionConfig) (VideoPrismVisionConfig 모델)
- **vilt** -- [ViltConfig](/docs/transformers/v5.23.0/ko/model_doc/vilt#transformers.ViltConfig) (ViltConfig 모델)
- **vipllava** -- [VipLlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaConfig) (VipLlavaConfig 모델)
- **vision-encoder-decoder** -- [VisionEncoderDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/vision-encoder-decoder#transformers.VisionEncoderDecoderConfig) (VisionEncoderDecoderConfig 모델)
- **vision-text-dual-encoder** -- [VisionTextDualEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/vision-text-dual-encoder#transformers.VisionTextDualEncoderConfig) (VisionTextDualEncoderConfig 모델)
- **visual_bert** -- [VisualBertConfig](/docs/transformers/v5.23.0/ko/model_doc/visual_bert#transformers.VisualBertConfig) (VisualBertConfig 모델)
- **vit** -- [ViTConfig](/docs/transformers/v5.23.0/ko/model_doc/vit#transformers.ViTConfig) (ViTConfig 모델)
- **vit_mae** -- [ViTMAEConfig](/docs/transformers/v5.23.0/ko/model_doc/vit_mae#transformers.ViTMAEConfig) (ViTMAEConfig 모델)
- **vit_msn** -- [ViTMSNConfig](/docs/transformers/v5.23.0/ko/model_doc/vit_msn#transformers.ViTMSNConfig) (ViTMSNConfig 모델)
- **vitdet** -- [VitDetConfig](/docs/transformers/v5.23.0/ko/model_doc/vitdet#transformers.VitDetConfig) (VitDetConfig 모델)
- **vitmatte** -- [VitMatteConfig](/docs/transformers/v5.23.0/ko/model_doc/vitmatte#transformers.VitMatteConfig) (VitMatteConfig 모델)
- **vitpose** -- [VitPoseConfig](/docs/transformers/v5.23.0/ko/model_doc/vitpose#transformers.VitPoseConfig) (VitPoseConfig 모델)
- **vitpose_backbone** -- `VitPoseBackboneConfig` (VitPoseBackboneConfig 모델)
- **vits** -- [VitsConfig](/docs/transformers/v5.23.0/ko/model_doc/vits#transformers.VitsConfig) (VitsConfig 모델)
- **vivit** -- [VivitConfig](/docs/transformers/v5.23.0/ko/model_doc/vivit#transformers.VivitConfig) (VivitConfig 모델)
- **vjepa2** -- [VJEPA2Config](/docs/transformers/v5.23.0/ko/model_doc/vjepa2#transformers.VJEPA2Config) (VJEPA2Config 모델)
- **voxtral** -- [VoxtralConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralConfig) (VoxtralConfig 모델)
- **voxtral_encoder** -- [VoxtralEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralEncoderConfig) (VoxtralEncoderConfig 모델)
- **voxtral_realtime** -- [VoxtralRealtimeConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeConfig) (VoxtralRealtimeConfig 모델)
- **voxtral_realtime_encoder** -- [VoxtralRealtimeEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeEncoderConfig) (VoxtralRealtimeEncoderConfig 모델)
- **voxtral_realtime_text** -- [VoxtralRealtimeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeTextConfig) (VoxtralRealtimeTextConfig 모델)
- **wav2vec2** -- [Wav2Vec2Config](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Config) (Wav2Vec2Config 모델)
- **wav2vec2-bert** -- [Wav2Vec2BertConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertConfig) (Wav2Vec2BertConfig 모델)
- **wav2vec2-conformer** -- [Wav2Vec2ConformerConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerConfig) (Wav2Vec2ConformerConfig 모델)
- **wavlm** -- [WavLMConfig](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMConfig) (WavLMConfig 모델)
- **whisper** -- [WhisperConfig](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperConfig) (WhisperConfig 모델)
- **xclip** -- [XCLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/xclip#transformers.XCLIPConfig) (XCLIPConfig 모델)
- **xclip_text_model** -- [XCLIPTextConfig](/docs/transformers/v5.23.0/ko/model_doc/xclip#transformers.XCLIPTextConfig) (XCLIPTextConfig 모델)
- **xclip_vision_model** -- [XCLIPVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/xclip#transformers.XCLIPVisionConfig) (XCLIPVisionConfig 모델)
- **xcodec** -- [XcodecConfig](/docs/transformers/v5.23.0/ko/model_doc/xcodec#transformers.XcodecConfig) (XcodecConfig 모델)
- **xcodec2** -- [Xcodec2Config](/docs/transformers/v5.23.0/ko/model_doc/xcodec2#transformers.Xcodec2Config) (Xcodec2Config 모델)
- **xglm** -- [XGLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xglm#transformers.XGLMConfig) (XGLMConfig 모델)
- **xlm** -- [XLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMConfig) (XLMConfig 모델)
- **xlm-roberta** -- [XLMRobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaConfig) (XLMRobertaConfig 모델)
- **xlm-roberta-xl** -- [XLMRobertaXLConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLConfig) (XLMRobertaXLConfig 모델)
- **xlnet** -- [XLNetConfig](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetConfig) (XLNetConfig 모델)
- **xlstm** -- [xLSTMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlstm#transformers.xLSTMConfig) (xLSTMConfig 모델)
- **xmod** -- [XmodConfig](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodConfig) (XmodConfig 모델)
- **yolos** -- [YolosConfig](/docs/transformers/v5.23.0/ko/model_doc/yolos#transformers.YolosConfig) (YolosConfig 모델)
- **yoso** -- [YosoConfig](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoConfig) (YosoConfig 모델)
- **youtu** -- [YoutuConfig](/docs/transformers/v5.23.0/ko/model_doc/youtu#transformers.YoutuConfig) (YoutuConfig 모델)
- **zamba** -- [ZambaConfig](/docs/transformers/v5.23.0/ko/model_doc/zamba#transformers.ZambaConfig) (ZambaConfig 모델)
- **zamba2** -- [Zamba2Config](/docs/transformers/v5.23.0/ko/model_doc/zamba2#transformers.Zamba2Config) (Zamba2Config 모델)
- **zaya** -- [ZayaConfig](/docs/transformers/v5.23.0/ko/model_doc/zaya#transformers.ZayaConfig) (ZayaConfig 모델)
- **zoedepth** -- [ZoeDepthConfig](/docs/transformers/v5.23.0/ko/model_doc/zoedepth#transformers.ZoeDepthConfig) (ZoeDepthConfig 모델)

예:

```python
>>> from transformers import AutoConfig

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-uncased")

>>> # Download configuration from huggingface.co (user-uploaded) and cache.
>>> config = AutoConfig.from_pretrained("dbmdz/bert-base-german-cased")

>>> # If configuration file is in a directory (e.g., was saved using *save_pretrained('./test/saved_model/')*).
>>> config = AutoConfig.from_pretrained("./test/bert_saved_model/")

>>> # Load a specific configuration file.
>>> config = AutoConfig.from_pretrained("./test/bert_saved_model/my_configuration.json")

>>> # Change some config attributes when loading a pretrained config.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-uncased", output_attentions=True, foo=False)
>>> config.output_attentions
True

>>> config, unused_kwargs = AutoConfig.from_pretrained(
...     "google-bert/bert-base-uncased", output_attentions=True, foo=False, return_unused_kwargs=True
... )
>>> config.output_attentions
True

>>> unused_kwargs
{'foo': False}
```

#### register[[transformers.AutoConfig.register]]

```python
register(model_type, config, exist_ok = False)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/configuration_auto.py#L455)

**매개변수:**

model_type (`str`) : "bert" 또는 "gpt"과 같은 모델 유형.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 등록할 config입니다.

이 클래스에 대한 새 구성을 등록합니다.

## AutoTokenizer[[transformers.AutoTokenizer]]

#### transformers.AutoTokenizer[[transformers.AutoTokenizer]]

```python
transformers.AutoTokenizer()
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/tokenization_auto.py#L659)

이것은 [AutoTokenizer.from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoTokenizer.from_pretrained) 클래스 메서드로 생성될 때 라이브러리의 토크나이저 클래스 중 하나로 인스턴스화되는 범용 토크나이저 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_pretrained[[transformers.AutoTokenizer.from_pretrained]]

```python
from_pretrained(pretrained_model_name_or_path, *inputs, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/tokenization_auto.py#L673)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 정의된 tokenizer의 *model id*인 문자열. - tokenizer에 필요한 어휘 파일이 포함된 *디렉터리*의 경로. 예를 들어 [save_pretrained()](/docs/transformers/v5.23.0/ko/internal/tokenization_utils#transformers.PreTrainedTokenizerBase.save_pretrained) 메서드를 사용하여 저장한 경우, 예: `./my_model_directory/`. - tokenizer가 단일 어휘 파일만 요구하는 경우(예: Bert 또는 XLNet)에만 해당되는 단일 저장 어휘 파일의 경로, 예: `./my_model_directory/vocab.txt`. (모든 파생 클래스에 적용되지 않음)

inputs (추가 위치 인자, *선택 사항*) : Tokenizer의 `__init__()` 메서드로 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 인스턴스화할 토크나이저 클래스를 결정하는 데 사용되는 설정 객체입니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 모델 가중치 및 구성 파일의 (재)다운로드를 강제하고, 캐시된 버전이 존재하는 경우 이를 덮어쓸지 여부.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

subfolder (`str`, *선택 사항*) : 관련 파일이 huggingface.co의 모델 저장소 내 하위 폴더에 있는 경우(예: facebook/rag-token-base), 여기에 지정합니다.

tokenizer_type (`str`, *선택 사항*) : 로드할 Tokenizer 유형.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

kwargs (추가 키워드 인자, *선택 사항*) : Tokenizer의 `__init__()` 메서드에 전달됩니다. `bos_token`, `eos_token`, `unk_token`, `sep_token`, `pad_token`, `cls_token`, `mask_token`, `additional_special_tokens`와 같은 특수 토큰을 설정하는 데 사용할 수 있습니다. 자세한 내용은 `__init__()`의 매개변수를 참조하세요.

pretrained model 어휘에서 라이브러리의 토크나이저 클래스 중 하나를 인스턴스화합니다.

인스턴스화할 토크나이저 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **EvollaModel** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (EvollaConfig 모델)
- **aimv2** -- [CLIPTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTokenizer) (Aimv2Config 모델)
- **albert** -- [AlbertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertTokenizer) (AlbertConfig 모델)
- **align** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (AlignConfig 모델)
- **aria** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (AriaConfig 모델)
- **audioflamingo3** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (AudioFlamingo3Config 모델)
- **aya_vision** -- [CohereTokenizer](/docs/transformers/v5.23.0/ko/model_doc/cohere#transformers.CohereTokenizer) (AyaVisionConfig 모델)
- **bark** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (BarkConfig 모델)
- **bart** -- [RobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (BartConfig 모델)
- **beam** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (BeamConfig 모델)
- **bert** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (BertConfig 모델)
- **bert-generation** -- [BertGenerationTokenizer](/docs/transformers/v5.23.0/ko/model_doc/bert-generation#transformers.BertGenerationTokenizer) (BertGenerationConfig 모델)
- **big_bird** -- [BigBirdTokenizer](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdTokenizer) (BigBirdConfig 모델)
- **bigbird_pegasus** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (BigBirdPegasusConfig 모델)
- **biogpt** -- [BioGptTokenizer](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptTokenizer) (BioGptConfig 모델)
- **blenderbot** -- [BlenderbotTokenizer](/docs/transformers/v5.23.0/ko/model_doc/blenderbot#transformers.BlenderbotTokenizer) (BlenderbotConfig 모델)
- **blenderbot-small** -- [BlenderbotSmallTokenizer](/docs/transformers/v5.23.0/ko/model_doc/blenderbot-small#transformers.BlenderbotSmallTokenizer) (BlenderbotSmallConfig 모델)
- **blip** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (BlipConfig 모델)
- **blip-2** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (Blip2Config 모델)
- **bridgetower** -- [RobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (BridgeTowerConfig 모델)
- **bros** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (BrosConfig 모델)
- **camembert** -- [CamembertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertTokenizer) (CamembertConfig 모델)
- **canary** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (CanaryConfig 모델)
- **canine** -- [CanineTokenizer](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineTokenizer) (CanineConfig 모델)
- **chinese_clip** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (ChineseCLIPConfig 모델)
- **clap** -- [RobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (ClapConfig 모델)
- **clip** -- [CLIPTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTokenizer) (CLIPConfig 모델)
- **clipseg** -- [CLIPTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTokenizer) (CLIPSegConfig 모델)
- **clvp** -- [ClvpTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clvp#transformers.ClvpTokenizer) (ClvpConfig 모델)
- **codegen** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (CodeGenConfig 모델)
- **cohere** -- [CohereTokenizer](/docs/transformers/v5.23.0/ko/model_doc/cohere#transformers.CohereTokenizer) (CohereConfig 모델)
- **cohere2** -- [CohereTokenizer](/docs/transformers/v5.23.0/ko/model_doc/cohere#transformers.CohereTokenizer) (Cohere2Config 모델)
- **cohere_asr** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (CohereAsrConfig 모델)
- **colqwen2** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (ColQwen2Config 모델)
- **convbert** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (ConvBertConfig 모델)
- **cosmos3_omni** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Cosmos3OmniConfig 모델)
- **cpmant** -- [CpmAntTokenizer](/docs/transformers/v5.23.0/ko/model_doc/cpmant#transformers.CpmAntTokenizer) (CpmAntConfig 모델)
- **ctrl** -- [CTRLTokenizer](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLTokenizer) (CTRLConfig 모델)
- **data2vec-audio** -- [Wav2Vec2CTCTokenizer](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2CTCTokenizer) (Data2VecAudioConfig 모델)
- **data2vec-text** -- [RobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (Data2VecTextConfig 모델)
- **dbrx** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (DbrxConfig 모델)
- **deberta** -- [DebertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaTokenizer) (DebertaConfig 모델)
- **deberta-v2** -- [DebertaV2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2Tokenizer) (DebertaV2Config 모델)
- **dia** -- [DiaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaTokenizer) (DiaConfig 모델)
- **diffusion_gemma** -- [GemmaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaTokenizer) (DiffusionGemmaConfig 모델)
- **distilbert** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (DistilBertConfig 모델)
- **dots3_note** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Dots3NoteConfig 모델)
- **dpr** -- [DPRQuestionEncoderTokenizer](/docs/transformers/v5.23.0/ko/model_doc/dpr#transformers.DPRQuestionEncoderTokenizer) (DPRConfig 모델)
- **electra** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (ElectraConfig 모델)
- **emu3** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (Emu3Config 모델)
- **ernie** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (ErnieConfig 모델)
- **ernie4_5** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Ernie4_5Config 모델)
- **ernie4_5_moe** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Ernie4_5_MoeConfig 모델)
- **esm** -- [EsmTokenizer](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmTokenizer) (EsmConfig 모델)
- **esmc** -- [EsmcTokenizer](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcTokenizer) (EsmcConfig 모델)
- **esmfold2** -- [EsmcTokenizer](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcTokenizer) (EsmFold2Config 모델)
- **falcon_mamba** -- [GPTNeoXTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXTokenizer) (FalconMambaConfig 모델)
- **fastspeech2_conformer** -- `None` (FastSpeech2ConformerConfig 모델)
- **flaubert** -- [FlaubertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertTokenizer) (FlaubertConfig 모델)
- **flava** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (FlavaConfig 모델)
- **flex_olmo** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (FlexOlmoConfig 모델)
- **florence2** -- [BartTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (Florence2Config 모델)
- **fnet** -- [FNetTokenizer](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetTokenizer) (FNetConfig 모델)
- **fsmt** -- [FSMTTokenizer](/docs/transformers/v5.23.0/ko/model_doc/fsmt#transformers.FSMTTokenizer) (FSMTConfig 모델)
- **fun_asr_nano** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (FunAsrNanoConfig 모델)
- **funnel** -- [FunnelTokenizer](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelTokenizer) (FunnelConfig 모델)
- **fuyu** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (FuyuConfig 모델)
- **gemma** -- [GemmaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaTokenizer) (GemmaConfig 모델)
- **gemma2** -- [GemmaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaTokenizer) (Gemma2Config 모델)
- **gemma3** -- [GemmaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaTokenizer) (Gemma3Config 모델)
- **gemma3_text** -- [GemmaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaTokenizer) (Gemma3TextConfig 모델)
- **gemma3n** -- [GemmaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaTokenizer) (Gemma3nConfig 모델)
- **gemma3n_text** -- [GemmaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaTokenizer) (Gemma3nTextConfig 모델)
- **git** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (GitConfig 모델)
- **glm** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (GlmConfig 모델)
- **glm4** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Glm4Config 모델)
- **glm4_moe** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Glm4MoeConfig 모델)
- **glm4_moe_lite** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Glm4MoeLiteConfig 모델)
- **glm4v** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Glm4vConfig 모델)
- **glm4v_moe** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Glm4vMoeConfig 모델)
- **glm_image** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (GlmImageConfig 모델)
- **glmasr** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (GlmAsrConfig 모델)
- **got_ocr2** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (GotOcr2Config 모델)
- **gpt-sw3** -- [GPTSw3Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt-sw3#transformers.GPTSw3Tokenizer) (GPT2Config 모델)
- **gpt2** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (GPT2Config 모델)
- **gpt_bigcode** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (GPTBigCodeConfig 모델)
- **gpt_neo** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (GPTNeoConfig 모델)
- **gpt_neox** -- [GPTNeoXTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXTokenizer) (GPTNeoXConfig 모델)
- **gpt_neox_japanese** -- [GPTNeoXJapaneseTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox_japanese#transformers.GPTNeoXJapaneseTokenizer) (GPTNeoXJapaneseConfig 모델)
- **gptj** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (GPTJConfig 모델)
- **gradium_phonon_tts** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (GradiumPhononTTSConfig 모델)
- **granite** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (GraniteConfig 모델)
- **granite_speech5_ctc** -- [ParakeetTokenizer](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetTokenizer) (GraniteSpeech5CTCConfig 모델)
- **granitemoe** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (GraniteMoeConfig 모델)
- **granitemoehybrid** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (GraniteMoeHybridConfig 모델)
- **granitemoeshared** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (GraniteMoeSharedConfig 모델)
- **grounding-dino** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (GroundingDinoConfig 모델)
- **groupvit** -- [CLIPTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTokenizer) (GroupViTConfig 모델)
- **gte** -- [XLMRobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaTokenizer) (GteConfig 모델)
- **hubert** -- [Wav2Vec2CTCTokenizer](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2CTCTokenizer) (HubertConfig 모델)
- **hunyuan_vl** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (HunYuanVLConfig 모델)
- **hyperclovax** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (HyperCLOVAXConfig 모델)
- **hyperclovax_vision_v2** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (HyperCLOVAXVisionV2Config 모델)
- **ibert** -- [RobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (IBertConfig 모델)
- **idefics** -- [LlamaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaTokenizer) (IdeficsConfig 모델)
- **idefics2** -- [LlamaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaTokenizer) (Idefics2Config 모델)
- **instructblip** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (InstructBlipConfig 모델)
- **instructblipvideo** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (InstructBlipVideoConfig 모델)
- **internlm3** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (InternLM3Config 모델)
- **internvl** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (InternVLConfig 모델)
- **iquestcoder** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (IQuestCoderConfig 모델)
- **jais2** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (Jais2Config 모델)
- **jamba** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (JambaConfig 모델)
- **jina_embeddings_v3** -- [XLMRobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaTokenizer) (JinaEmbeddingsV3Config 모델)
- **kimi_k25** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Kimi_K25Config 모델)
- **kimi_linear** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (KimiLinearConfig 모델)
- **kosmos-2** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Kosmos2Config 모델)
- **lasr_ctc** -- [LasrTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrTokenizer) (LasrCTCConfig 모델)
- **lasr_encoder** -- [LasrTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrTokenizer) (LasrEncoderConfig 모델)
- **layoutlm** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (LayoutLMConfig 모델)
- **layoutlmv2** -- [LayoutLMv2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2Tokenizer) (LayoutLMv2Config 모델)
- **layoutlmv3** -- [LayoutLMv3Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3Tokenizer) (LayoutLMv3Config 모델)
- **layoutxlm** -- [LayoutXLMTokenizer](/docs/transformers/v5.23.0/ko/model_doc/layoutxlm#transformers.LayoutXLMTokenizer) (LayoutXLMConfig 모델)
- **led** -- [LEDTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (LEDConfig 모델)
- **lighton_ocr** -- [Qwen2TokenizerFast](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (LightOnOcrConfig 모델)
- **lilt** -- [RobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (LiltConfig 모델)
- **llava** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (LlavaConfig 모델)
- **llava_next** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (LlavaNextConfig 모델)
- **longformer** -- [RobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (LongformerConfig 모델)
- **luke** -- [LukeTokenizer](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeTokenizer) (LukeConfig 모델)
- **lxmert** -- [LxmertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (LxmertConfig 모델)
- **m2m_100** -- [M2M100Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/m2m_100#transformers.M2M100Tokenizer) (M2M100Config 모델)
- **mamba** -- [GPTNeoXTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXTokenizer) (MambaConfig 모델)
- **mamba2** -- [GPTNeoXTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXTokenizer) (Mamba2Config 모델)
- **marian** -- [MarianTokenizer](/docs/transformers/v5.23.0/ko/model_doc/marian#transformers.MarianTokenizer) (MarianConfig 모델)
- **markuplm** -- [MarkupLMTokenizer](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMTokenizer) (MarkupLMConfig 모델)
- **mbart** -- [MBartTokenizer](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartTokenizer) (MBartConfig 모델)
- **megatron-bert** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (MegatronBertConfig 모델)
- **metaclip_2** -- [XLMRobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaTokenizer) (MetaClip2Config 모델)
- **mgp-str** -- [MgpstrTokenizer](/docs/transformers/v5.23.0/ko/model_doc/mgp-str#transformers.MgpstrTokenizer) (MgpstrConfig 모델)
- **mimo_v2_flash** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (MiMoV2FlashConfig 모델)
- **minicpm3** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (MiniCPM3Config 모델)
- **minicpm_ocr** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (MiniCPMOcrConfig 모델)
- **minicpmv4_6** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (MiniCPMV4_6Config 모델)
- **minicpmv4_7** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (MiniCPMV4_7Config 모델)
- **minimax_m2** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (MiniMaxM2Config 모델)
- **ministral** -- [MistralCommonBackend](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralCommonBackend) (MinistralConfig 모델)
- **ministral3** -- [MistralCommonBackend](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralCommonBackend) (Ministral3Config 모델)
- **mistral** -- [MistralCommonBackend](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralCommonBackend) (MistralConfig 모델)
- **mistral3** -- [MistralCommonBackend](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralCommonBackend) (Mistral3Config 모델)
- **mistral_large4** -- [MistralCommonBackend](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralCommonBackend) (MistralLarge4Config 모델)
- **mixtral** -- [MistralCommonBackend](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralCommonBackend) (MixtralConfig 모델)
- **mm-grounding-dino** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (MMGroundingDinoConfig 모델)
- **mobilebert** -- [MobileBertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (MobileBertConfig 모델)
- **modernbert** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (ModernBertConfig 모델)
- **modernbert-decoder** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (ModernBertDecoderConfig 모델)
- **molmo2** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Molmo2Config 모델)
- **moss_transcribe_diarize** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (MossTranscribeDiarizeConfig 모델)
- **mpnet** -- [MPNetTokenizer](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetTokenizer) (MPNetConfig 모델)
- **mpt** -- [GPTNeoXTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXTokenizer) (MptConfig 모델)
- **mra** -- [RobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (MraConfig 모델)
- **mt5** -- [T5Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Tokenizer) (MT5Config 모델)
- **musicgen** -- [T5Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Tokenizer) (MusicgenConfig 모델)
- **musicgen_melody** -- [T5Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Tokenizer) (MusicgenMelodyConfig 모델)
- **mvp** -- [MvpTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (MvpConfig 모델)
- **nemo_speechlm** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Nemotron3_5TranscribeConfig 모델)
- **nemotron** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (NemotronConfig 모델)
- **nemotron3_5_asr** -- [ParakeetTokenizer](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetTokenizer) (Nemotron3_5AsrConfig 모델)
- **nemotron_asr_streaming** -- [ParakeetTokenizer](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetTokenizer) (NemotronAsrStreamingConfig 모델)
- **nllb-moe** -- [NllbTokenizer](/docs/transformers/v5.23.0/ko/model_doc/nllb#transformers.NllbTokenizer) (NllbMoeConfig 모델)
- **nomic_bert** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (NomicBertConfig 모델)
- **nougat** -- [NougatTokenizer](/docs/transformers/v5.23.0/ko/model_doc/nougat#transformers.NougatTokenizer) (NougatConfig 모델)
- **nystromformer** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (NystromformerConfig 모델)
- **olmo** -- [GPTNeoXTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXTokenizer) (OlmoConfig 모델)
- **olmo2** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Olmo2Config 모델)
- **olmo3** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Olmo3Config 모델)
- **olmo_hybrid** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (OlmoHybridConfig 모델)
- **olmoe** -- [GPTNeoXTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXTokenizer) (OlmoeConfig 모델)
- **omdet-turbo** -- [CLIPTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTokenizer) (OmDetTurboConfig 모델)
- **oneformer** -- [CLIPTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTokenizer) (OneFormerConfig 모델)
- **openai-gpt** -- [OpenAIGPTTokenizer](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTTokenizer) (OpenAIGPTConfig 모델)
- **opt** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (OPTConfig 모델)
- **ovis2** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Ovis2Config 모델)
- **owlv2** -- [CLIPTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTokenizer) (Owlv2Config 모델)
- **owlvit** -- [CLIPTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTokenizer) (OwlViTConfig 모델)
- **paddleocr_vl** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (PaddleOCRVLConfig 모델)
- **parakeet_ctc** -- [ParakeetTokenizer](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetTokenizer) (ParakeetCTCConfig 모델)
- **parakeet_rnnt** -- [ParakeetTokenizer](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetTokenizer) (ParakeetRNNTConfig 모델)
- **parakeet_tdt** -- [ParakeetTokenizer](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetTokenizer) (ParakeetTDTConfig 모델)
- **pegasus** -- [PegasusTokenizer](/docs/transformers/v5.23.0/ko/model_doc/pegasus#transformers.PegasusTokenizer) (PegasusConfig 모델)
- **pegasus_x** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (PegasusXConfig 모델)
- **perceiver** -- [PerceiverTokenizer](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverTokenizer) (PerceiverConfig 모델)
- **persimmon** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (PersimmonConfig 모델)
- **phi** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (PhiConfig 모델)
- **phi3** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Phi3Config 모델)
- **phimoe** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (PhimoeConfig 모델)
- **pix2struct** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (Pix2StructConfig 모델)
- **pixtral** -- [MistralCommonBackend](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralCommonBackend) (PixtralVisionConfig 모델)
- **plbart** -- [PLBartTokenizer](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartTokenizer) (PLBartConfig 모델)
- **pocket_tts** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (PocketTTSConfig 모델)
- **pp_formulanet** -- [NougatTokenizer](/docs/transformers/v5.23.0/ko/model_doc/nougat#transformers.NougatTokenizer) (PPFormulaNetConfig 모델)
- **prophetnet** -- [ProphetNetTokenizer](/docs/transformers/v5.23.0/ko/model_doc/prophetnet#transformers.ProphetNetTokenizer) (ProphetNetConfig 모델)
- **qianfan_ocr** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (QianfanOCRConfig 모델)
- **qwen2** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen2Config 모델)
- **qwen2_5_omni** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen2_5OmniConfig 모델)
- **qwen2_5_vl** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen2_5_VLConfig 모델)
- **qwen2_audio** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen2AudioConfig 모델)
- **qwen2_moe** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen2MoeConfig 모델)
- **qwen2_vl** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen2VLConfig 모델)
- **qwen3** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen3Config 모델)
- **qwen3_5** -- [Qwen3_5Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Tokenizer) (Qwen3_5Config 모델)
- **qwen3_5_moe** -- [Qwen3_5Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Tokenizer) (Qwen3_5MoeConfig 모델)
- **qwen3_5_moe_text** -- [Qwen3_5Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Tokenizer) (Qwen3_5MoeTextConfig 모델)
- **qwen3_5_text** -- [Qwen3_5Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Tokenizer) (Qwen3_5TextConfig 모델)
- **qwen3_asr** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen3ASRConfig 모델)
- **qwen3_moe** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen3MoeConfig 모델)
- **qwen3_next** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen3NextConfig 모델)
- **qwen3_omni_moe** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen3OmniMoeConfig 모델)
- **qwen3_vl** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen3VLConfig 모델)
- **qwen3_vl_moe** -- [Qwen2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (Qwen3VLMoeConfig 모델)
- **qwen4_exp** -- [Qwen3_5Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Tokenizer) (Qwen4ExpConfig 모델)
- **rag** -- [RagTokenizer](/docs/transformers/v5.23.0/ko/model_doc/rag#transformers.RagTokenizer) (RagConfig 모델)
- **recurrent_gemma** -- [GemmaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaTokenizer) (RecurrentGemmaConfig 모델)
- **reformer** -- [ReformerTokenizer](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerTokenizer) (ReformerConfig 모델)
- **rembert** -- [RemBertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertTokenizer) (RemBertConfig 모델)
- **roberta** -- [RobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (RobertaConfig 모델)
- **roberta-prelayernorm** -- [RobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.RobertaTokenizer) (RobertaPreLayerNormConfig 모델)
- **roc_bert** -- [RoCBertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertTokenizer) (RoCBertConfig 모델)
- **roformer** -- [RoFormerTokenizer](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerTokenizer) (RoFormerConfig 모델)
- **rwkv** -- [GPTNeoXTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXTokenizer) (RwkvConfig 모델)
- **sam3** -- [CLIPTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTokenizer) (Sam3Config 모델)
- **sam3_video** -- [CLIPTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTokenizer) (Sam3VideoConfig 모델)
- **seamless_m4t** -- [SeamlessM4TTokenizer](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TTokenizer) (SeamlessM4TConfig 모델)
- **seamless_m4t_v2** -- [SeamlessM4TTokenizer](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TTokenizer) (SeamlessM4Tv2Config 모델)
- **shieldgemma2** -- [GemmaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaTokenizer) (ShieldGemma2Config 모델)
- **siglip** -- [SiglipTokenizer](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipTokenizer) (SiglipConfig 모델)
- **siglip2** -- [Siglip2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2Tokenizer) (Siglip2Config 모델)
- **smolvlm** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (SmolVLMConfig 모델)
- **speech_to_text** -- [Speech2TextTokenizer](/docs/transformers/v5.23.0/ko/model_doc/speech_to_text#transformers.Speech2TextTokenizer) (Speech2TextConfig 모델)
- **speecht5** -- [SpeechT5Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/speecht5#transformers.SpeechT5Tokenizer) (SpeechT5Config 모델)
- **splinter** -- [SplinterTokenizer](/docs/transformers/v5.23.0/ko/model_doc/splinter#transformers.SplinterTokenizer) (SplinterConfig 모델)
- **squeezebert** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (SqueezeBertConfig 모델)
- **stablelm** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (StableLmConfig 모델)
- **starcoder2** -- [GPT2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Tokenizer) (Starcoder2Config 모델)
- **switch_transformers** -- [T5Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Tokenizer) (SwitchTransformersConfig 모델)
- **t5** -- [T5Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Tokenizer) (T5Config 모델)
- **t5gemma** -- [GemmaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaTokenizer) (T5GemmaConfig 모델)
- **tapas** -- [TapasTokenizer](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasTokenizer) (TapasConfig 모델)
- **tipsv2** -- [Tipsv2Tokenizer](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2Tokenizer) (Tipsv2Config 모델)
- **trocr** -- [XLMRobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaTokenizer) (TrOCRConfig 모델)
- **tvp** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (TvpConfig 모델)
- **udop** -- [UdopTokenizer](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopTokenizer) (UdopConfig 모델)
- **umt5** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (UMT5Config 모델)
- **unispeech** -- [Wav2Vec2CTCTokenizer](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2CTCTokenizer) (UniSpeechConfig 모델)
- **unispeech-sat** -- [Wav2Vec2CTCTokenizer](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2CTCTokenizer) (UniSpeechSatConfig 모델)
- **vibevoice** -- [Qwen2TokenizerFast](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Tokenizer) (VibeVoiceConfig 모델)
- **videoprism** -- [VideoPrismTokenizer](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismTokenizer) (VideoPrismConfig 모델)
- **vilt** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (ViltConfig 모델)
- **vipllava** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (VipLlavaConfig 모델)
- **visual_bert** -- [BertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.BertTokenizer) (VisualBertConfig 모델)
- **vits** -- [VitsTokenizer](/docs/transformers/v5.23.0/ko/model_doc/vits#transformers.VitsTokenizer) (VitsConfig 모델)
- **voxtral** -- [MistralCommonBackend](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralCommonBackend) (VoxtralConfig 모델)
- **voxtral_realtime** -- [MistralCommonBackend](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralCommonBackend) (VoxtralRealtimeConfig 모델)
- **wav2vec2** -- [Wav2Vec2CTCTokenizer](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2CTCTokenizer) (Wav2Vec2Config 모델)
- **wav2vec2-bert** -- [Wav2Vec2CTCTokenizer](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2CTCTokenizer) (Wav2Vec2BertConfig 모델)
- **wav2vec2-conformer** -- [Wav2Vec2CTCTokenizer](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2CTCTokenizer) (Wav2Vec2ConformerConfig 모델)
- **whisper** -- [WhisperTokenizer](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperTokenizer) (WhisperConfig 모델)
- **xclip** -- [CLIPTokenizer](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTokenizer) (XCLIPConfig 모델)
- **xglm** -- [XGLMTokenizer](/docs/transformers/v5.23.0/ko/model_doc/xglm#transformers.XGLMTokenizer) (XGLMConfig 모델)
- **xlm** -- [XLMTokenizer](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMTokenizer) (XLMConfig 모델)
- **xlm-roberta** -- [XLMRobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaTokenizer) (XLMRobertaConfig 모델)
- **xlm-roberta-xl** -- [TokenizersBackend](/docs/transformers/v5.23.0/ko/main_classes/tokenizer#transformers.TokenizersBackend) (XLMRobertaXLConfig 모델)
- **xlnet** -- [XLNetTokenizer](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetTokenizer) (XLNetConfig 모델)
- **xlstm** -- [GPTNeoXTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXTokenizer) (xLSTMConfig 모델)
- **xmod** -- [XLMRobertaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaTokenizer) (XmodConfig 모델)
- **yoso** -- [AlbertTokenizer](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertTokenizer) (YosoConfig 모델)
- **zaya** -- [GemmaTokenizer](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaTokenizer) (ZayaConfig 모델)

예:

```python
>>> from transformers import AutoTokenizer

>>> # Download vocabulary from huggingface.co and cache.
>>> tokenizer = AutoTokenizer.from_pretrained("google-bert/bert-base-uncased")

>>> # Download vocabulary from huggingface.co (user-uploaded) and cache.
>>> tokenizer = AutoTokenizer.from_pretrained("dbmdz/bert-base-german-cased")

>>> # If vocabulary files are in a directory (e.g. tokenizer was saved using *save_pretrained('./test/saved_model/')*)
>>> # tokenizer = AutoTokenizer.from_pretrained("./test/bert_saved_model/")

>>> # Download vocabulary from huggingface.co and define model-specific arguments
>>> tokenizer = AutoTokenizer.from_pretrained("FacebookAI/roberta-base", add_prefix_space=True)
```

#### register[[transformers.AutoTokenizer.register]]

```python
register(config_class, tokenizer_class = None, slow_tokenizer_class = None, fast_tokenizer_class = None, exist_ok = False)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/tokenization_auto.py#L1034)

**매개변수:**

config_class ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 등록할 모델에 해당하는 설정입니다.

tokenizer_class : 등록할 토크나이저 클래스 (V5 - 권장 매개변수).

slow_tokenizer_class : (비추천) 등록할 느린 토크나이저.

fast_tokenizer_class : (비추천) 등록할 빠른 토크나이저입니다.

이 매핑에 새로운 tokenizer를 등록합니다.

## AutoFeatureExtractor[[transformers.AutoFeatureExtractor]]

#### transformers.AutoFeatureExtractor[[transformers.AutoFeatureExtractor]]

```python
transformers.AutoFeatureExtractor()
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/feature_extraction_auto.py#L232)

이것은 [AutoFeatureExtractor.from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoFeatureExtractor.from_pretrained) 클래스 메서드로 생성될 때 라이브러리의 feature extractor 클래스 중 하나로 인스턴스화되는 범용 feature extractor 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_pretrained[[transformers.AutoFeatureExtractor.from_pretrained]]

```python
from_pretrained(pretrained_model_name_or_path, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/feature_extraction_auto.py#L246)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습된 feature_extractor의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/feature_extractor#transformers.FeatureExtractionMixin.save_pretrained) 메서드를 사용하여 저장된 feature extractor 파일이 포함된 *디렉터리*의 경로, 예: `./my_model_directory/`. - 저장된 feature extractor JSON *파일*의 경로, 예: `./my_model_directory/preprocessor_config.json`.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 feature extractor를 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : feature extractor 파일을 (재)다운로드를 강제하고, 캐시된 버전이 존재하는 경우 이를 덮어쓸지 여부.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}.` 각 요청에 프록시가 사용됩니다.

token (`str` 또는 *bool*, *선택*) : 원격 파일에 대한 HTTP bearer 인증으로 사용할 토큰입니다. `True`인 경우, `hf auth login` 실행 시 생성된 토큰(`~/.huggingface`에 저장됨)을 사용합니다.

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

return_unused_kwargs (`bool`, *선택 사항*, 기본값 `False`) : `False`인 경우, 이 함수는 최종 feature extractor 객체만 반환합니다. `True`인 경우, 이 함수는 `Tuple(feature_extractor, unused_kwargs)`를 반환하며, 여기서 *unused_kwargs*는 feature extractor 속성이 아닌 키/값 쌍으로 구성된 사전입니다. 즉, `kwargs` 중 `feature_extractor`를 업데이트하는 데 사용되지 않고 무시되는 부분입니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

kwargs (`dict[str, Any]`, *선택*) : kwargs의 값 중 feature extractor 속성에 해당하는 키는 로드된 값을 덮어쓰는 데 사용됩니다. feature extractor 속성이 *아닌* 키/값 쌍에 대한 동작은 `return_unused_kwargs` 키워드 매개변수로 제어됩니다.

…

인스턴스화할 feature extractor 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된)에 따라 선택되며, 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백됩니다:

- **audio-spectrogram-transformer** -- [ASTFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/audio-spectrogram-transformer#transformers.ASTFeatureExtractor) (ASTConfig 모델)
- **audioflamingo3** -- [WhisperFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperFeatureExtractor) (AudioFlamingo3Config 모델)
- **canary** -- [ParakeetFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetFeatureExtractor) (CanaryConfig 모델)
- **clap** -- [ClapFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/clap#transformers.ClapFeatureExtractor) (ClapConfig 모델)
- **clvp** -- [ClvpFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/clvp#transformers.ClvpFeatureExtractor) (ClvpConfig 모델)
- **cohere_asr** -- [CohereAsrFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/cohere_asr#transformers.CohereAsrFeatureExtractor) (CohereAsrConfig 모델)
- **csm** -- [EncodecFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/encodec#transformers.EncodecFeatureExtractor) (CsmConfig 모델)
- **dac** -- [DacFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/dac#transformers.DacFeatureExtractor) (DacConfig 모델)
- **data2vec-audio** -- [Wav2Vec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2FeatureExtractor) (Data2VecAudioConfig 모델)
- **dia** -- [DiaFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaFeatureExtractor) (DiaConfig 모델)
- **dots3_note** -- [Dots3NoteFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteFeatureExtractor) (Dots3NoteConfig 모델)
- **embedding_gemma2** -- [Gemma4AudioFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4AudioFeatureExtractor) (EmbeddingGemma2Config 모델)
- **encodec** -- [EncodecFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/encodec#transformers.EncodecFeatureExtractor) (EncodecConfig 모델)
- **fun_asr_nano** -- [FunAsrNanoFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoFeatureExtractor) (FunAsrNanoConfig 모델)
- **gemma3n** -- [Gemma3nAudioFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nAudioFeatureExtractor) (Gemma3nConfig 모델)
- **gemma4** -- [Gemma4AudioFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4AudioFeatureExtractor) (Gemma4Config 모델)
- **gemma4_unified** -- [Gemma4UnifiedAudioFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedAudioFeatureExtractor) (Gemma4UnifiedConfig 모델)
- **glmasr** -- [WhisperFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperFeatureExtractor) (GlmAsrConfig 모델)
- **gradium_phonon_tts** -- [PocketTTSFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts#transformers.PocketTTSFeatureExtractor) (GradiumPhononTTSConfig 모델)
- **granite_speech** -- [GraniteSpeechFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechFeatureExtractor) (GraniteSpeechConfig 모델)
- **granite_speech5_ctc** -- [GraniteSpeech5FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5FeatureExtractor) (GraniteSpeech5CTCConfig 모델)
- **granite_speech5_encoder** -- [GraniteSpeech5FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5FeatureExtractor) (GraniteSpeech5EncoderConfig 모델)
- **granite_speech_plus** -- [GraniteSpeechFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechFeatureExtractor) (GraniteSpeechPlusConfig 모델)
- **higgs_audio_v2_tokenizer** -- [DacFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/dac#transformers.DacFeatureExtractor) (HiggsAudioV2TokenizerConfig 모델)
- **hubert** -- [Wav2Vec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2FeatureExtractor) (HubertConfig 모델)
- **inkling_mm_model** -- [InklingFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingFeatureExtractor) (InklingConfig 모델)
- **kyutai_speech_to_text** -- [KyutaiSpeechToTextFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextFeatureExtractor) (KyutaiSpeechToTextConfig 모델)
- **lasr_ctc** -- [LasrFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrFeatureExtractor) (LasrCTCConfig 모델)
- **lasr_encoder** -- [LasrFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrFeatureExtractor) (LasrEncoderConfig 모델)
- **mimi** -- [EncodecFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/encodec#transformers.EncodecFeatureExtractor) (MimiConfig 모델)
- **moonshine** -- [Wav2Vec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2FeatureExtractor) (MoonshineConfig 모델)
- **moshi** -- [EncodecFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/encodec#transformers.EncodecFeatureExtractor) (MoshiConfig 모델)
- **moss_transcribe_diarize** -- [MossTranscribeDiarizeFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeFeatureExtractor) (MossTranscribeDiarizeConfig 모델)
- **muse_spark** -- [MuseSparkFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkFeatureExtractor) (MuseSparkConfig 모델)
- **musicgen** -- [EncodecFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/encodec#transformers.EncodecFeatureExtractor) (MusicgenConfig 모델)
- **musicgen_melody** -- [MusicgenMelodyFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/musicgen_melody#transformers.MusicgenMelodyFeatureExtractor) (MusicgenMelodyConfig 모델)
- **nemo_speechlm** -- [NemotronAsrStreamingFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingFeatureExtractor) (Nemotron3_5TranscribeConfig 모델)
- **nemotron3_5_asr** -- [NemotronAsrStreamingFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingFeatureExtractor) (Nemotron3_5AsrConfig 모델)
- **nemotron3_diarization** -- [NemotronAsrStreamingFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingFeatureExtractor) (Nemotron3DiarizationConfig 모델)
- **nemotron_asr_streaming** -- [NemotronAsrStreamingFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingFeatureExtractor) (NemotronAsrStreamingConfig 모델)
- **nemotron_asr_streaming_encoder** -- [NemotronAsrStreamingFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingFeatureExtractor) (NemotronAsrStreamingEncoderConfig 모델)
- **neucodec** -- [NeuCodecFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/neucodec#transformers.NeuCodecFeatureExtractor) (NeuCodecConfig 모델)
- **parakeet_ctc** -- [ParakeetFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetFeatureExtractor) (ParakeetCTCConfig 모델)
- **parakeet_encoder** -- [ParakeetFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetFeatureExtractor) (ParakeetEncoderConfig 모델)
- **parakeet_rnnt** -- [ParakeetFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetFeatureExtractor) (ParakeetRNNTConfig 모델)
- **parakeet_tdt** -- [ParakeetFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetFeatureExtractor) (ParakeetTDTConfig 모델)
- **pe_audio** -- [PeAudioFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/pe_audio#transformers.PeAudioFeatureExtractor) (PeAudioConfig 모델)
- **pe_audio_video** -- [PeAudioFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/pe_audio#transformers.PeAudioFeatureExtractor) (PeAudioVideoConfig 모델)
- **phi4_multimodal** -- [Phi4MultimodalFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalFeatureExtractor) (Phi4MultimodalConfig 모델)
- **pocket_tts** -- [PocketTTSFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts#transformers.PocketTTSFeatureExtractor) (PocketTTSConfig 모델)
- **pocket_tts_audio_tokenizer** -- [EncodecFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/encodec#transformers.EncodecFeatureExtractor) (PocketTTSAudioTokenizerConfig 모델)
- **pop2piano** -- [Pop2PianoFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/pop2piano#transformers.Pop2PianoFeatureExtractor) (Pop2PianoConfig 모델)
- **qwen2_5_omni** -- [WhisperFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperFeatureExtractor) (Qwen2_5OmniConfig 모델)
- **qwen2_audio** -- [WhisperFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperFeatureExtractor) (Qwen2AudioConfig 모델)
- **qwen3_asr** -- [Qwen3ASRFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRFeatureExtractor) (Qwen3ASRConfig 모델)
- **qwen3_omni_moe** -- [WhisperFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperFeatureExtractor) (Qwen3OmniMoeConfig 모델)
- **seamless_m4t** -- [SeamlessM4TFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TFeatureExtractor) (SeamlessM4TConfig 모델)
- **seamless_m4t_v2** -- [SeamlessM4TFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TFeatureExtractor) (SeamlessM4Tv2Config 모델)
- **sew** -- [Wav2Vec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2FeatureExtractor) (SEWConfig 모델)
- **sew-d** -- [Wav2Vec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2FeatureExtractor) (SEWDConfig 모델)
- **speech_to_text** -- [Speech2TextFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/speech_to_text#transformers.Speech2TextFeatureExtractor) (Speech2TextConfig 모델)
- **speecht5** -- [SpeechT5FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/speecht5#transformers.SpeechT5FeatureExtractor) (SpeechT5Config 모델)
- **unispeech** -- [Wav2Vec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2FeatureExtractor) (UniSpeechConfig 모델)
- **unispeech-sat** -- [Wav2Vec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2FeatureExtractor) (UniSpeechSatConfig 모델)
- **univnet** -- [UnivNetFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/univnet#transformers.UnivNetFeatureExtractor) (UnivNetConfig 모델)
- **vibevoice** -- [VibeVoiceAcousticTokenizerFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerFeatureExtractor) (VibeVoiceConfig 모델)
- **vibevoice_acoustic_tokenizer** -- [VibeVoiceAcousticTokenizerFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerFeatureExtractor) (VibeVoiceAcousticTokenizerConfig 모델)
- **vibevoice_asr** -- [VibeVoiceAcousticTokenizerFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerFeatureExtractor) (VibeVoiceAsrConfig 모델)
- **voxtral** -- [WhisperFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperFeatureExtractor) (VoxtralConfig 모델)
- **voxtral_realtime** -- [VoxtralRealtimeFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeFeatureExtractor) (VoxtralRealtimeConfig 모델)
- **wav2vec2** -- [Wav2Vec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2FeatureExtractor) (Wav2Vec2Config 모델)
- **wav2vec2-bert** -- [Wav2Vec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2FeatureExtractor) (Wav2Vec2BertConfig 모델)
- **wav2vec2-conformer** -- [Wav2Vec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2FeatureExtractor) (Wav2Vec2ConformerConfig 모델)
- **wavlm** -- [Wav2Vec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2FeatureExtractor) (WavLMConfig 모델)
- **whisper** -- [WhisperFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperFeatureExtractor) (WhisperConfig 모델)
- **xcodec** -- [DacFeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/dac#transformers.DacFeatureExtractor) (XcodecConfig 모델)
- **xcodec2** -- [Xcodec2FeatureExtractor](/docs/transformers/v5.23.0/ko/model_doc/xcodec2#transformers.Xcodec2FeatureExtractor) (Xcodec2Config 모델)

비공개 모델을 사용하려면 `token=True`를 전달해야 합니다.

예:

```python
>>> from transformers import AutoFeatureExtractor

>>> # Download feature extractor from huggingface.co and cache.
>>> feature_extractor = AutoFeatureExtractor.from_pretrained("facebook/wav2vec2-base-960h")

>>> # If feature extractor files are in a directory (e.g. feature extractor was saved using *save_pretrained('./test/saved_model/')*)
>>> # feature_extractor = AutoFeatureExtractor.from_pretrained("./test/saved_model/")
```

#### register[[transformers.AutoFeatureExtractor.register]]

```python
register(config_class, feature_extractor_class, exist_ok = False)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/feature_extraction_auto.py#L383)

**매개변수:**

config_class ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 등록할 모델에 해당하는 설정입니다.

feature_extractor_class (`FeatureExtractorMixin`) : 등록할 feature extractor입니다.

이 클래스에 새로운 feature extractor를 등록합니다.

## AutoImageProcessor[[transformers.AutoImageProcessor]]

#### transformers.AutoImageProcessor[[transformers.AutoImageProcessor]]

```python
transformers.AutoImageProcessor()
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/image_processing_auto.py#L499)

이것은 [AutoImageProcessor.from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoImageProcessor.from_pretrained) 클래스 메서드로 생성될 때 라이브러리의 이미지 프로세서 클래스 중 하나로 인스턴스화되는 범용 이미지 프로세서 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_pretrained[[transformers.AutoImageProcessor.from_pretrained]]

```python
from_pretrained(pretrained_model_name_or_path, *inputs, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/image_processing_auto.py#L513)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다: - huggingface.co의 모델 저장소에 호스팅된 사전 학습된 image_processor의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/internal/image_processing_utils#transformers.ImageProcessingMixin.save_pretrained) 메서드를 사용하여 저장된 image processor 파일이 포함된 *디렉터리*의 경로, 예: `./my_model_directory/`. - 저장된 image processor JSON *파일*의 경로, 예: `./my_model_directory/preprocessor_config.json`.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 이미지 프로세서가 캐시될 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 이미지가 존재하는 경우 캐시된 버전을 덮어쓰고 이미지 프로세서 파일을 (재)다운로드하도록 강제할지 여부.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}.` 각 요청에 프록시가 사용됩니다.

token (`str` 또는 *bool*, *선택*) : 원격 파일에 대한 HTTP bearer 인증으로 사용할 토큰입니다. `True`인 경우, `hf auth login` 실행 시 생성된 토큰(`~/.huggingface`에 저장됨)을 사용합니다.

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

use_fast (`bool`, *선택 사항*, 기본값 `False`) : **비추천**: 대신 `backend="torchvision"`을 사용하세요. 이 매개변수는 하위 호환성을 위해 유지됩니다. 해당 모델에 대해 지원되는 경우, 빠른 torchvision 기반 이미지 프로세서를 사용합니다. 해당 모델에 대해 빠른 이미지 프로세서가 사용 가능하지 않으면, 대신 일반적인 numpy 기반 이미지 프로세서가 반환됩니다.

backend (`str`, *선택 사항*, 기본값 `None`) : 이미지 처리에 사용할 백엔드입니다. 다음 값 중 하나를 사용할 수 있습니다: - `None`: 사용 가능한 최적의 백엔드를 자동으로 선택합니다 (torchvision이 사용 가능한 경우 torchvision, 그렇지 않으면 pil) - `"torchvision"`: Torchvision 백엔드를 사용합니다 (GPU 가속, 더 빠름) - `"pil"`: PIL 백엔드를 사용합니다 (이식성, CPU 전용) - `register()` 메서드를 통해 등록된 사용자 정의 백엔드 이름

return_unused_kwargs (`bool`, *선택 사항*, 기본값 `False`) : `False`인 경우, 이 함수는 최종 image processor 객체만 반환합니다. `True`인 경우, 이 함수는 `Tuple(image_processor, unused_kwargs)`를 반환하며, 여기서 *unused_kwargs*는 image processor 속성이 아닌 키/값 쌍으로 구성된 사전입니다. 즉, `image_processor`를 업데이트하는 데 사용되지 않고 무시되는 `kwargs`의 부분입니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

image_processor_filename (`str`, *선택 사항*, 기본값 `"config.json"`) : 이미지 프로세서 설정에 사용할 모델 디렉터리의 파일 이름입니다.

kwargs (`dict[str, Any]`, *선택*) : kwargs의 값 중 이미지 프로세서 속성에 해당하는 키는 로드된 값을 덮어쓰는 데 사용됩니다. 키가 이미지 프로세서 속성이 *아닌* 키/값 쌍에 대한 동작은 `return_unused_kwargs` 키워드 매개변수로 제어됩니다.

…

인스턴스화할 이미지 프로세서 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드됨)에 따라 선택되며, 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백됩니다:

- **aimv2** -- `{'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'}` (Aimv2Config 모델)
- **aimv2_vision_model** -- `{'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'}` (Aimv2VisionConfig 모델)
- **align** -- `{'torchvision': 'EfficientNetImageProcessor', 'pil': 'EfficientNetImageProcessorPil'}` (AlignConfig 모델)
- **altclip** -- `{'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'}` (AltCLIPConfig 모델)
- **aria** -- `{'pil': 'AriaImageProcessorPil', 'torchvision': 'AriaImageProcessor'}` (AriaConfig 모델)
- **aya_vision** -- `{'torchvision': 'GotOcr2ImageProcessor', 'pil': 'GotOcr2ImageProcessorPil'}` (AyaVisionConfig 모델)
- **beit** -- `{'pil': 'BeitImageProcessorPil', 'torchvision': 'BeitImageProcessor'}` (BeitConfig 모델)
- **bit** -- `{'pil': 'BitImageProcessorPil', 'torchvision': 'BitImageProcessor'}` (BitConfig 모델)
- **blip** -- `{'pil': 'BlipImageProcessorPil', 'torchvision': 'BlipImageProcessor'}` (BlipConfig 모델)
- **blip-2** -- `{'torchvision': 'BlipImageProcessor', 'pil': 'BlipImageProcessorPil'}` (Blip2Config 모델)
- **bridgetower** -- `{'pil': 'BridgeTowerImageProcessorPil', 'torchvision': 'BridgeTowerImageProcessor'}` (BridgeTowerConfig 모델)
- **chameleon** -- `{'pil': 'ChameleonImageProcessorPil', 'torchvision': 'ChameleonImageProcessor'}` (ChameleonConfig 모델)
- **chinese_clip** -- `{'pil': 'ChineseCLIPImageProcessorPil', 'torchvision': 'ChineseCLIPImageProcessor'}` (ChineseCLIPConfig 모델)
- **chmv2** -- `{'torchvision': 'CHMv2ImageProcessor'}` (CHMv2Config 모델)
- **clip** -- `{'pil': 'CLIPImageProcessorPil', 'torchvision': 'CLIPImageProcessor'}` (CLIPConfig 모델)
- **clipseg** -- `{'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'}` (CLIPSegConfig 모델)
- **cohere2_vision** -- `{'torchvision': 'Cohere2VisionImageProcessor'}` (Cohere2VisionConfig 모델)
- **cohere_compass** -- `{'pil': 'CohereCompassImageProcessorPil', 'torchvision': 'CohereCompassImageProcessor'}` (CohereCompassConfig 모델)
- **colpali** -- `{'torchvision': 'SiglipImageProcessor', 'pil': 'SiglipImageProcessorPil'}` (ColPaliConfig 모델)
- **colqwen2** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (ColQwen2Config 모델)
- **conditional_detr** -- `{'pil': 'ConditionalDetrImageProcessorPil', 'torchvision': 'ConditionalDetrImageProcessor'}` (ConditionalDetrConfig 모델)
- **convnext** -- `{'pil': 'ConvNextImageProcessorPil', 'torchvision': 'ConvNextImageProcessor'}` (ConvNextConfig 모델)
- **convnextv2** -- `{'torchvision': 'ConvNextImageProcessor', 'pil': 'ConvNextImageProcessorPil'}` (ConvNextV2Config 모델)
- **cosmos3_edge** -- `{'pil': 'Cosmos3EdgeImageProcessorPil', 'torchvision': 'Cosmos3EdgeImageProcessor'}` (Cosmos3EdgeConfig 모델)
- **cosmos3_omni** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (Cosmos3OmniConfig 모델)
- **cvt** -- `{'torchvision': 'ConvNextImageProcessor', 'pil': 'ConvNextImageProcessorPil'}` (CvtConfig 모델)
- **data2vec-vision** -- `{'torchvision': 'BeitImageProcessor', 'pil': 'BeitImageProcessorPil'}` (Data2VecVisionConfig 모델)
- **deepseek_ocr2** -- `{'pil': 'DeepseekOcr2ImageProcessorPil', 'torchvision': 'DeepseekOcr2ImageProcessor'}` (DeepseekOcr2Config 모델)
- **deepseek_vl** -- `{'pil': 'DeepseekVLImageProcessorPil', 'torchvision': 'DeepseekVLImageProcessor'}` (DeepseekVLConfig 모델)
- **deepseek_vl_hybrid** -- `{'pil': 'DeepseekVLHybridImageProcessorPil', 'torchvision': 'DeepseekVLHybridImageProcessor'}` (DeepseekVLHybridConfig 모델)
- **deformable_detr** -- `{'pil': 'DeformableDetrImageProcessorPil', 'torchvision': 'DeformableDetrImageProcessor'}` (DeformableDetrConfig 모델)
- **deimv2** -- `{'torchvision': 'RTDetrImageProcessor', 'pil': 'RTDetrImageProcessorPil'}` (Deimv2Config 모델)
- **deit** -- `{'pil': 'DeiTImageProcessorPil', 'torchvision': 'DeiTImageProcessor'}` (DeiTConfig 모델)
- **depth_anything** -- `{'torchvision': 'DPTImageProcessor', 'pil': 'DPTImageProcessorPil'}` (DepthAnythingConfig 모델)
- **depth_pro** -- `{'torchvision': 'DepthProImageProcessor'}` (DepthProConfig 모델)
- **detr** -- `{'pil': 'DetrImageProcessorPil', 'torchvision': 'DetrImageProcessor'}` (DetrConfig 모델)
- **dinat** -- `{'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'}` (DinatConfig 모델)
- **dinov2** -- `{'torchvision': 'BitImageProcessor', 'pil': 'BitImageProcessorPil'}` (Dinov2Config 모델)
- **dinov3_vit** -- `{'torchvision': 'DINOv3ViTImageProcessor'}` (DINOv3ViTConfig 모델)
- **donut-swin** -- `{'torchvision': 'DonutImageProcessor', 'pil': 'DonutImageProcessorPil'}` (DonutSwinConfig 모델)
- **dots3_note** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (Dots3NoteConfig 모델)
- **dpt** -- `{'pil': 'DPTImageProcessorPil', 'torchvision': 'DPTImageProcessor'}` (DPTConfig 모델)
- **edgetam** -- `{'torchvision': 'Sam2ImageProcessor'}` (EdgeTamConfig 모델)
- **efficientloftr** -- `{'pil': 'EfficientLoFTRImageProcessorPil', 'torchvision': 'EfficientLoFTRImageProcessor'}` (EfficientLoFTRConfig 모델)
- **efficientnet** -- `{'pil': 'EfficientNetImageProcessorPil', 'torchvision': 'EfficientNetImageProcessor'}` (EfficientNetConfig 모델)
- **embedding_gemma2** -- `{'pil': 'Gemma4ImageProcessorPil', 'torchvision': 'Gemma4ImageProcessor'}` (EmbeddingGemma2Config 모델)
- **emu3** -- `{'pil': 'Emu3ImageProcessor'}` (Emu3Config 모델)
- **eomt** -- `{'pil': 'EomtImageProcessorPil', 'torchvision': 'EomtImageProcessor'}` (EomtConfig 모델)
- **eomt_dinov3** -- `{'torchvision': 'EomtImageProcessor', 'pil': 'EomtImageProcessorPil'}` (EomtDinov3Config 모델)
- **ernie4_5_vl_moe** -- `{'pil': 'Ernie4_5_VLMoeImageProcessorPil', 'torchvision': 'Ernie4_5_VLMoeImageProcessor'}` (Ernie4_5_VLMoeConfig 모델)
- **exaone4_5** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (Exaone4_5_Config 모델)
- **flava** -- `{'pil': 'FlavaImageProcessorPil', 'torchvision': 'FlavaImageProcessor'}` (FlavaConfig 모델)
- **florence2** -- `{'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'}` (Florence2Config 모델)
- **focalnet** -- `{'torchvision': 'BitImageProcessor', 'pil': 'BitImageProcessorPil'}` (FocalNetConfig 모델)
- **fuyu** -- `{'pil': 'FuyuImageProcessorPil', 'torchvision': 'FuyuImageProcessor'}` (FuyuConfig 모델)
- **gemma3** -- `{'pil': 'Gemma3ImageProcessorPil', 'torchvision': 'Gemma3ImageProcessor'}` (Gemma3Config 모델)
- **gemma3n** -- `{'torchvision': 'SiglipImageProcessor', 'pil': 'SiglipImageProcessorPil'}` (Gemma3nConfig 모델)
- **gemma4** -- `{'pil': 'Gemma4ImageProcessorPil', 'torchvision': 'Gemma4ImageProcessor'}` (Gemma4Config 모델)
- **gemma4_unified** -- `{'torchvision': 'Gemma4UnifiedImageProcessor'}` (Gemma4UnifiedConfig 모델)
- **git** -- `{'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'}` (GitConfig 모델)
- **glm46v** -- `{'pil': 'Glm46VImageProcessorPil', 'torchvision': 'Glm46VImageProcessor'}` (Glm46VConfig 모델)
- **glm4v** -- `{'pil': 'Glm4vImageProcessorPil', 'torchvision': 'Glm4vImageProcessor'}` (Glm4vConfig 모델)
- **glm5_next** -- `{'pil': 'Glm5NextImageProcessorPil', 'torchvision': 'Glm5NextImageProcessor'}` (Glm5NextConfig 모델)
- **glm_image** -- `{'pil': 'GlmImageImageProcessorPil', 'torchvision': 'GlmImageImageProcessor'}` (GlmImageConfig 모델)
- **glmga** -- `{'pil': 'GlmgaImageProcessorPil', 'torchvision': 'GlmgaImageProcessor'}` (GlmgaConfig 모델)
- **glpn** -- `{'pil': 'GLPNImageProcessorPil', 'torchvision': 'GLPNImageProcessor'}` (GLPNConfig 모델)
- **got_ocr2** -- `{'pil': 'GotOcr2ImageProcessorPil', 'torchvision': 'GotOcr2ImageProcessor'}` (GotOcr2Config 모델)
- **granite4_vision** -- `{'torchvision': 'LlavaNextImageProcessor', 'pil': 'LlavaNextImageProcessorPil'}` (Granite4VisionConfig 모델)
- **granite_docling** -- `{'pil': 'GraniteDoclingImageProcessorPil', 'torchvision': 'GraniteDoclingImageProcessor'}` (GraniteDoclingConfig 모델)
- **grounding-dino** -- `{'pil': 'GroundingDinoImageProcessorPil', 'torchvision': 'GroundingDinoImageProcessor'}` (GroundingDinoConfig 모델)
- **groupvit** -- `{'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'}` (GroupViTConfig 모델)
- **hiera** -- `{'torchvision': 'BitImageProcessor', 'pil': 'BitImageProcessorPil'}` (HieraConfig 모델)
- **hunyuan_vl** -- `{'pil': 'HunYuanVLImageProcessorPil', 'torchvision': 'HunYuanVLImageProcessor'}` (HunYuanVLConfig 모델)
- **hyperclovax_vision_v2** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (HyperCLOVAXVisionV2Config 모델)
- **idefics** -- `{'pil': 'IdeficsImageProcessorPil', 'torchvision': 'IdeficsImageProcessor'}` (IdeficsConfig 모델)
- **idefics2** -- `{'pil': 'Idefics2ImageProcessorPil', 'torchvision': 'Idefics2ImageProcessor'}` (Idefics2Config 모델)
- **idefics3** -- `{'pil': 'Idefics3ImageProcessorPil', 'torchvision': 'Idefics3ImageProcessor'}` (Idefics3Config 모델)
- **ijepa** -- `{'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'}` (IJepaConfig 모델)
- **imagegpt** -- `{'pil': 'ImageGPTImageProcessorPil', 'torchvision': 'ImageGPTImageProcessor'}` (ImageGPTConfig 모델)
- **inkling2** -- `{'torchvision': 'Inkling2ImageProcessor'}` (Inkling2Config 모델)
- **inkling_mm_model** -- `{'torchvision': 'InklingImageProcessor'}` (InklingConfig 모델)
- **instructblip** -- `{'torchvision': 'BlipImageProcessor', 'pil': 'BlipImageProcessorPil'}` (InstructBlipConfig 모델)
- **internvl** -- `{'torchvision': 'GotOcr2ImageProcessor', 'pil': 'GotOcr2ImageProcessorPil'}` (InternVLConfig 모델)
- **janus** -- `{'pil': 'JanusImageProcessorPil', 'torchvision': 'JanusImageProcessor'}` (JanusConfig 모델)
- **kimi_k25** -- `{'torchvision': 'Kimi_K25ImageProcessor'}` (Kimi_K25Config 모델)
- **kosmos-2** -- `{'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'}` (Kosmos2Config 모델)
- **kosmos-2.5** -- `{'torchvision': 'Kosmos2_5ImageProcessor', 'pil': 'Kosmos2_5ImageProcessorPil'}` (Kosmos2_5Config 모델)
- **layoutlmv2** -- `{'pil': 'LayoutLMv2ImageProcessorPil', 'torchvision': 'LayoutLMv2ImageProcessor'}` (LayoutLMv2Config 모델)
- **layoutlmv3** -- `{'pil': 'LayoutLMv3ImageProcessorPil', 'torchvision': 'LayoutLMv3ImageProcessor'}` (LayoutLMv3Config 모델)
- **layoutxlm** -- `{'torchvision': 'LayoutLMv2ImageProcessor', 'pil': 'LayoutLMv2ImageProcessorPil'}` (LayoutXLMConfig 모델)
- **levit** -- `{'pil': 'LevitImageProcessorPil', 'torchvision': 'LevitImageProcessor'}` (LevitConfig 모델)
- **lfm2_vl** -- `{'torchvision': 'Lfm2VlImageProcessor'}` (Lfm2VlConfig 모델)
- **lfm3_vl** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (Lfm3VlConfig 모델)
- **lightglue** -- `{'pil': 'LightGlueImageProcessorPil', 'torchvision': 'LightGlueImageProcessor'}` (LightGlueConfig 모델)
- **lighton_ocr** -- `{'torchvision': 'PixtralImageProcessor', 'pil': 'PixtralImageProcessorPil'}` (LightOnOcrConfig 모델)
- **llama4** -- `{'torchvision': 'Llama4ImageProcessor'}` (Llama4Config 모델)
- **llava** -- `{'pil': 'LlavaImageProcessorPil', 'torchvision': 'LlavaImageProcessor'}` (LlavaConfig 모델)
- **llava_next** -- `{'pil': 'LlavaNextImageProcessorPil', 'torchvision': 'LlavaNextImageProcessor'}` (LlavaNextConfig 모델)
- **llava_next_video** -- `{'torchvision': 'LlavaNextImageProcessor', 'pil': 'LlavaNextImageProcessorPil'}` (LlavaNextVideoConfig 모델)
- **llava_onevision** -- `{'pil': 'LlavaOnevisionImageProcessorPil', 'torchvision': 'LlavaOnevisionImageProcessor'}` (LlavaOnevisionConfig 모델)
- **lw_detr** -- `{'torchvision': 'DeformableDetrImageProcessor', 'pil': 'DeformableDetrImageProcessorPil'}` (LwDetrConfig 모델)
- **mask2former** -- `{'pil': 'Mask2FormerImageProcessorPil', 'torchvision': 'Mask2FormerImageProcessor'}` (Mask2FormerConfig 모델)
- **maskformer** -- `{'pil': 'MaskFormerImageProcessorPil', 'torchvision': 'MaskFormerImageProcessor'}` (MaskFormerConfig 모델)
- **metaclip_2** -- `{'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'}` (MetaClip2Config 모델)
- **mgp-str** -- `{'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'}` (MgpstrConfig 모델)
- **minicpm_ocr** -- `{'torchvision': 'MiniCPMOcrImageProcessor'}` (MiniCPMOcrConfig 모델)
- **minicpmv4_6** -- `{'pil': 'MiniCPMV4_6ImageProcessorPil', 'torchvision': 'MiniCPMV4_6ImageProcessor'}` (MiniCPMV4_6Config 모델)
- **minicpmv4_7** -- `{'pil': 'MiniCPMV4_6ImageProcessorPil', 'torchvision': 'MiniCPMV4_6ImageProcessor'}` (MiniCPMV4_7Config 모델)
- **minimax_m3_vl** -- `{'torchvision': 'MiniMaxM3VLImageProcessor'}` (MiniMaxM3VLConfig 모델)
- **mistral3** -- `{'torchvision': 'PixtralImageProcessor', 'pil': 'PixtralImageProcessorPil'}` (Mistral3Config 모델)
- **mistral_large4** -- `{'torchvision': 'PixtralImageProcessor', 'pil': 'PixtralImageProcessorPil'}` (MistralLarge4Config 모델)
- **mlcd** -- `{'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'}` (MLCDVisionConfig 모델)
- **mllama** -- `{'pil': 'MllamaImageProcessorPil', 'torchvision': 'MllamaImageProcessor'}` (MllamaConfig 모델)
- **mm-grounding-dino** -- `{'torchvision': 'GroundingDinoImageProcessor', 'pil': 'GroundingDinoImageProcessorPil'}` (MMGroundingDinoConfig 모델)
- **mobilenet_v1** -- `{'pil': 'MobileNetV1ImageProcessorPil', 'torchvision': 'MobileNetV1ImageProcessor'}` (MobileNetV1Config 모델)
- **mobilenet_v2** -- `{'pil': 'MobileNetV2ImageProcessorPil', 'torchvision': 'MobileNetV2ImageProcessor'}` (MobileNetV2Config 모델)
- **mobilevit** -- `{'pil': 'MobileViTImageProcessorPil', 'torchvision': 'MobileViTImageProcessor'}` (MobileViTConfig 모델)
- **mobilevitv2** -- `{'torchvision': 'MobileViTImageProcessor', 'pil': 'MobileViTImageProcessorPil'}` (MobileViTV2Config 모델)
- **molmo2** -- `{'torchvision': 'Molmo2ImageProcessor'}` (Molmo2Config 모델)
- **muse_glimmer** -- `{'torchvision': 'MuseGlimmerImageProcessor'}` (MuseGlimmerConfig 모델)
- **muse_spark** -- `{'torchvision': 'MuseGlimmerImageProcessor'}` (MuseSparkConfig 모델)
- **nemotron_h_omni** -- `{'torchvision': 'NemotronH_Omni_Reasoning_V3ImageProcessor'}` (NemotronH_Omni_Reasoning_V3_Config 모델)
- **neomme** -- `{'torchvision': 'NeoMMEImageProcessor'}` (NeoMMEConfig 모델)
- **nougat** -- `{'pil': 'NougatImageProcessorPil', 'torchvision': 'NougatImageProcessor'}` (NougatConfig 모델)
- **omdet-turbo** -- `{'torchvision': 'DetrImageProcessor', 'pil': 'DetrImageProcessorPil'}` (OmDetTurboConfig 모델)
- **oneformer** -- `{'pil': 'OneFormerImageProcessorPil', 'torchvision': 'OneFormerImageProcessor'}` (OneFormerConfig 모델)
- **ovis2** -- `{'pil': 'Ovis2ImageProcessorPil', 'torchvision': 'Ovis2ImageProcessor'}` (Ovis2Config 모델)
- **owlv2** -- `{'pil': 'Owlv2ImageProcessorPil', 'torchvision': 'Owlv2ImageProcessor'}` (Owlv2Config 모델)
- **owlvit** -- `{'pil': 'OwlViTImageProcessorPil', 'torchvision': 'OwlViTImageProcessor'}` (OwlViTConfig 모델)
- **paddleocr_vl** -- `{'pil': 'PaddleOCRVLImageProcessorPil', 'torchvision': 'PaddleOCRVLImageProcessor'}` (PaddleOCRVLConfig 모델)
- **paligemma** -- `{'torchvision': 'SiglipImageProcessor', 'pil': 'SiglipImageProcessorPil'}` (PaliGemmaConfig 모델)
- **perceiver** -- `{'pil': 'PerceiverImageProcessorPil', 'torchvision': 'PerceiverImageProcessor'}` (PerceiverConfig 모델)
- **perception_lm** -- `{'torchvision': 'PerceptionLMImageProcessor'}` (PerceptionLMConfig 모델)
- **phi4_multimodal** -- `{'torchvision': 'Phi4MultimodalImageProcessor'}` (Phi4MultimodalConfig 모델)
- **pi0** -- `{'torchvision': 'PI0ImageProcessor'}` (PI0Config 모델)
- **pix2struct** -- `{'pil': 'Pix2StructImageProcessorPil', 'torchvision': 'Pix2StructImageProcessor'}` (Pix2StructConfig 모델)
- **pixio** -- `{'torchvision': 'BitImageProcessor', 'pil': 'BitImageProcessorPil'}` (PixioConfig 모델)
- **pixtral** -- `{'pil': 'PixtralImageProcessorPil', 'torchvision': 'PixtralImageProcessor'}` (PixtralVisionConfig 모델)
- **poolformer** -- `{'pil': 'PoolFormerImageProcessorPil', 'torchvision': 'PoolFormerImageProcessor'}` (PoolFormerConfig 모델)
- **pp_chart2table** -- `{'pil': 'PPChart2TableImageProcessorPil', 'torchvision': 'PPChart2TableImageProcessor'}` (PPChart2TableConfig 모델)
- **pp_doclayout_v2** -- `{'torchvision': 'PPDocLayoutV2ImageProcessor'}` (PPDocLayoutV2Config 모델)
- **pp_doclayout_v3** -- `{'torchvision': 'PPDocLayoutV3ImageProcessor'}` (PPDocLayoutV3Config 모델)
- **pp_doclayout_v4** -- `{'torchvision': 'PPDocLayoutV4ImageProcessor'}` (PPDocLayoutV4Config 모델)
- **pp_formulanet** -- `{'torchvision': 'PPFormulaNetImageProcessor'}` (PPFormulaNetConfig 모델)
- **pp_lcnet** -- `{'torchvision': 'PPLCNetImageProcessor'}` (PPLCNetConfig 모델)
- **pp_ocrv5_mobile_det** -- `{'torchvision': 'PPOCRV5ServerDetImageProcessor'}` (PPOCRV5MobileDetConfig 모델)
- **pp_ocrv5_mobile_rec** -- `{'torchvision': 'PPOCRV5ServerRecImageProcessor'}` (PPOCRV5MobileRecConfig 모델)
- **pp_ocrv5_server_det** -- `{'torchvision': 'PPOCRV5ServerDetImageProcessor'}` (PPOCRV5ServerDetConfig 모델)
- **pp_ocrv5_server_rec** -- `{'torchvision': 'PPOCRV5ServerRecImageProcessor'}` (PPOCRV5ServerRecConfig 모델)
- **pp_ocrv6_medium_det** -- `{'torchvision': 'PPOCRV5ServerDetImageProcessor'}` (PPOCRV6MediumDetConfig 모델)
- **pp_ocrv6_small_det** -- `{'torchvision': 'PPOCRV5ServerDetImageProcessor'}` (PPOCRV6SmallDetConfig 모델)
- **pp_ocrv6_small_rec** -- `{'torchvision': 'PPOCRV6SmallRecImageProcessor'}` (PPOCRV6SmallRecConfig 모델)
- **pp_ocrv6_tiny_rec** -- `{'torchvision': 'PPOCRV6SmallRecImageProcessor'}` (PPOCRV6TinyRecConfig 모델)
- **prompt_depth_anything** -- `{'pil': 'PromptDepthAnythingImageProcessorPil', 'torchvision': 'PromptDepthAnythingImageProcessor'}` (PromptDepthAnythingConfig 모델)
- **pvt** -- `{'pil': 'PvtImageProcessorPil', 'torchvision': 'PvtImageProcessor'}` (PvtConfig 모델)
- **pvt_v2** -- `{'torchvision': 'PvtImageProcessor', 'pil': 'PvtImageProcessorPil'}` (PvtV2Config 모델)
- **qianfan_ocr** -- `{'torchvision': 'GotOcr2ImageProcessor', 'pil': 'GotOcr2ImageProcessorPil'}` (QianfanOCRConfig 모델)
- **qwen2_5_omni** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (Qwen2_5OmniConfig 모델)
- **qwen2_5_vl** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (Qwen2_5_VLConfig 모델)
- **qwen2_vl** -- `{'pil': 'Qwen2VLImageProcessorPil', 'torchvision': 'Qwen2VLImageProcessor'}` (Qwen2VLConfig 모델)
- **qwen3_5** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (Qwen3_5Config 모델)
- **qwen3_5_moe** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (Qwen3_5MoeConfig 모델)
- **qwen3_omni_moe** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (Qwen3OmniMoeConfig 모델)
- **qwen3_vl** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (Qwen3VLConfig 모델)
- **qwen4_exp** -- `{'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'}` (Qwen4ExpConfig 모델)
- **regnet** -- `{'torchvision': 'ConvNextImageProcessor', 'pil': 'ConvNextImageProcessorPil'}` (RegNetConfig 모델)
- **resnet** -- `{'torchvision': 'ConvNextImageProcessor', 'pil': 'ConvNextImageProcessorPil'}` (ResNetConfig 모델)
- **rf_detr** -- `{'torchvision': 'RfDetrImageProcessor'}` (RfDetrConfig 모델)
- **rt_detr** -- `{'pil': 'RTDetrImageProcessorPil', 'torchvision': 'RTDetrImageProcessor'}` (RTDetrConfig 모델)
- **sam** -- `{'pil': 'SamImageProcessorPil', 'torchvision': 'SamImageProcessor'}` (SamConfig 모델)
- **sam2** -- `{'torchvision': 'Sam2ImageProcessor'}` (Sam2Config 모델)
- **sam2_video** -- `{'torchvision': 'Sam2ImageProcessor'}` (Sam2VideoConfig 모델)
- **sam3** -- `{'torchvision': 'Sam3ImageProcessor'}` (Sam3Config 모델)
- **sam3_lite_text** -- `{'torchvision': 'Sam3ImageProcessor'}` (Sam3LiteTextConfig 모델)
- **sam3_tracker** -- `{'torchvision': 'Sam3ImageProcessor'}` (Sam3TrackerConfig 모델)
- **sam3_tracker_video** -- `{'torchvision': 'Sam3ImageProcessor'}` (Sam3TrackerVideoConfig 모델)
- **sam3_video** -- `{'torchvision': 'Sam3ImageProcessor'}` (Sam3VideoConfig 모델)
- **sam_hq** -- `{'torchvision': 'SamImageProcessor', 'pil': 'SamImageProcessorPil'}` (SamHQConfig 모델)
- **sapiens2** -- `{'torchvision': 'Sapiens2ImageProcessor'}` (Sapiens2Config 모델)
- **segformer** -- `{'pil': 'SegformerImageProcessorPil', 'torchvision': 'SegformerImageProcessor'}` (SegformerConfig 모델)
- **seggpt** -- `{'pil': 'SegGptImageProcessorPil', 'torchvision': 'SegGptImageProcessor'}` (SegGptConfig 모델)
- **shieldgemma2** -- `{'torchvision': 'Gemma3ImageProcessor', 'pil': 'Gemma3ImageProcessorPil'}` (ShieldGemma2Config 모델)
- **siglip** -- `{'pil': 'SiglipImageProcessorPil', 'torchvision': 'SiglipImageProcessor'}` (SiglipConfig 모델)
- **siglip2** -- `{'pil': 'Siglip2ImageProcessorPil', 'torchvision': 'Siglip2ImageProcessor'}` (Siglip2Config 모델)
- **slanet** -- `{'torchvision': 'SLANeXtImageProcessor'}` (SLANetConfig 모델)
- **slanext** -- `{'torchvision': 'SLANeXtImageProcessor'}` (SLANeXtConfig 모델)
- **smolvlm** -- `{'pil': 'SmolVLMImageProcessorPil', 'torchvision': 'SmolVLMImageProcessor'}` (SmolVLMConfig 모델)
- **step3p7** -- `{'torchvision': 'Step3p7ImageProcessor'}` (Step3p7Config 모델)
- **superglue** -- `{'pil': 'SuperGlueImageProcessorPil', 'torchvision': 'SuperGlueImageProcessor'}` (SuperGlueConfig 모델)
- **superpoint** -- `{'pil': 'SuperPointImageProcessorPil', 'torchvision': 'SuperPointImageProcessor'}` (SuperPointConfig 모델)
- **swiftformer** -- `{'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'}` (SwiftFormerConfig 모델)
- **swin** -- `{'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'}` (SwinConfig 모델)
- **swin2sr** -- `{'pil': 'Swin2SRImageProcessorPil', 'torchvision': 'Swin2SRImageProcessor'}` (Swin2SRConfig 모델)
- **swinv2** -- `{'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'}` (Swinv2Config 모델)
- **t5gemma2** -- `{'torchvision': 'Gemma3ImageProcessor', 'pil': 'Gemma3ImageProcessorPil'}` (T5Gemma2Config 모델)
- **t5gemma2_encoder** -- `{'torchvision': 'Gemma3ImageProcessor', 'pil': 'Gemma3ImageProcessorPil'}` (T5Gemma2EncoderConfig 모델)
- **table-transformer** -- `{'torchvision': 'DetrImageProcessor', 'pil': 'DetrImageProcessorPil'}` (TableTransformerConfig 모델)
- **textnet** -- `{'pil': 'TextNetImageProcessorPil', 'torchvision': 'TextNetImageProcessor'}` (TextNetConfig 모델)
- **timesformer** -- `{'pil': 'VideoMAEImageProcessorPil', 'torchvision': 'VideoMAEImageProcessor'}` (TimesformerConfig 모델)
- **timm_wrapper** -- `{'pil': 'TimmWrapperImageProcessor'}` (TimmWrapperConfig 모델)
- **tipsv2** -- `{'torchvision': 'Tipsv2ImageProcessor'}` (Tipsv2Config 모델)
- **tipsv2_dpt** -- `{'torchvision': 'Tipsv2DptImageProcessor'}` (Tipsv2DptConfig 모델)
- **trocr** -- `{'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'}` (TrOCRConfig 모델)
- **tvp** -- `{'pil': 'TvpImageProcessorPil', 'torchvision': 'TvpImageProcessor'}` (TvpConfig 모델)
- **udop** -- `{'torchvision': 'LayoutLMv3ImageProcessor', 'pil': 'LayoutLMv3ImageProcessorPil'}` (UdopConfig 모델)
- **ultravbert** -- `{'torchvision': 'UltraVBertImageProcessor'}` (UltraVBertConfig 모델)
- **unlimited_ocr** -- `{'torchvision': 'UnlimitedOcrImageProcessor'}` (UnlimitedOcrConfig 모델)
- **upernet** -- `{'torchvision': 'SegformerImageProcessor', 'pil': 'SegformerImageProcessorPil'}` (UperNetConfig 모델)
- **uvdoc** -- `{'torchvision': 'UVDocImageProcessor'}` (UVDocConfig 모델)
- **video_llama_3** -- `{'pil': 'VideoLlama3ImageProcessorPil', 'torchvision': 'VideoLlama3ImageProcessor'}` (VideoLlama3Config 모델)
- **video_llava** -- `{'pil': 'VideoLlavaImageProcessor'}` (VideoLlavaConfig 모델)
- **videomae** -- `{'pil': 'VideoMAEImageProcessorPil', 'torchvision': 'VideoMAEImageProcessor'}` (VideoMAEConfig 모델)
- **vilt** -- `{'pil': 'ViltImageProcessorPil', 'torchvision': 'ViltImageProcessor'}` (ViltConfig 모델)
- **vipllava** -- `{'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'}` (VipLlavaConfig 모델)
- **vit** -- `{'pil': 'ViTImageProcessorPil', 'torchvision': 'ViTImageProcessor'}` (ViTConfig 모델)
- **vit_mae** -- `{'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'}` (ViTMAEConfig 모델)
- **vit_msn** -- `{'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'}` (ViTMSNConfig 모델)
- **vitmatte** -- `{'pil': 'VitMatteImageProcessorPil', 'torchvision': 'VitMatteImageProcessor'}` (VitMatteConfig 모델)
- **vitpose** -- `{'pil': 'VitPoseImageProcessorPil', 'torchvision': 'VitPoseImageProcessor'}` (VitPoseConfig 모델)
- **vivit** -- `{'torchvision': 'VivitImageProcessor'}` (VivitConfig 모델)
- **xclip** -- `{'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'}` (XCLIPConfig 모델)
- **yolos** -- `{'pil': 'YolosImageProcessorPil', 'torchvision': 'YolosImageProcessor'}` (YolosConfig 모델)
- **zoedepth** -- `{'pil': 'ZoeDepthImageProcessorPil', 'torchvision': 'ZoeDepthImageProcessor'}` (ZoeDepthConfig 모델)

비공개 모델을 사용하려면 `token=True`를 전달해야 합니다.

예:

```python
>>> from transformers import AutoImageProcessor

>>> # Download image processor from huggingface.co and cache.
>>> image_processor = AutoImageProcessor.from_pretrained("google/vit-base-patch16-224-in21k")

>>> # If image processor files are in a directory (e.g. image processor was saved using *save_pretrained('./test/saved_model/')*)
>>> # image_processor = AutoImageProcessor.from_pretrained("./test/saved_model/")
```

#### register[[transformers.AutoImageProcessor.register]]

```python
register(config_class, slow_image_processor_class: type | None = None, fast_image_processor_class: type | None = None, image_processor_classes: dict[str, type] | None = None, exist_ok: bool = False)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/image_processing_auto.py#L727)

**매개변수:**

config_class ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 등록할 모델에 해당하는 설정입니다.

slow_image_processor_class (`type`, *선택 사항*) : PIL 백엔드 이미지 프로세서 클래스(비추천, `image_processor_classes={"pil": ...}` 사용).

fast_image_processor_class (`type`, *선택 사항*) : Torchvision 백엔드 이미지 프로세서 클래스(비추천, `image_processor_classes={"torchvision": ...}` 사용).

image_processor_classes (`dict[str, type]`, *선택 사항*) : 백엔드 이름을 이미지 프로세서 클래스로 매핑하는 사전입니다. 사용자 정의 백엔드를 등록할 수 있습니다. 예: `{"pil": MyPilProcessor, "torchvision": MyTorchvisionProcessor, "custom": MyCustomProcessor}`

exist_ok (`bool`, *선택 사항*, 기본값 `False`) : `True`인 경우 기존 등록을 덮어쓰는 것을 허용합니다.

이 클래스에 새로운 이미지 프로세서를 등록합니다.

## AutoVideoProcessor[[transformers.AutoVideoProcessor]]

백엔드 선택, 사용자 정의 등록, 저장 및 재로딩에 대해서는 [Video processor backends](../main_classes/video_processor#video-processor-backends)를 참조하세요.

#### transformers.AutoVideoProcessor[[transformers.AutoVideoProcessor]]

```python
transformers.AutoVideoProcessor()
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/video_processing_auto.py#L266)

이것은 [AutoVideoProcessor.from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoVideoProcessor.from_pretrained) 클래스 메서드로 생성될 때 라이브러리의 비디오 프로세서 클래스 중 하나로 인스턴스화되는 범용 비디오 프로세서 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_pretrained[[transformers.AutoVideoProcessor.from_pretrained]]

```python
from_pretrained(pretrained_model_name_or_path, *inputs, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/video_processing_auto.py#L280)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습된 video_processor의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/video_processor#transformers.VideoProcessorMixin.save_pretrained) 메서드를 사용하여 저장된 video processor 파일이 포함된 *디렉터리*의 경로, 예: `./my_model_directory/`. - 저장된 video processor JSON *파일*의 경로, 예: `./my_model_directory/preprocessor_config.json`.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델의 video processor를 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 비디오 프로세서 파일을 (재)다운로드를 강제하고, 캐시된 버전이 존재하는 경우 이를 덮어쓸지 여부.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}.` 각 요청에 프록시가 사용됩니다.

token (`str` 또는 *bool*, *선택*) : 원격 파일에 대한 HTTP bearer 인증으로 사용할 토큰입니다. `True`인 경우, `hf auth login` 실행 시 생성된 토큰(`~/.huggingface`에 저장됨)을 사용합니다.

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

return_unused_kwargs (`bool`, *선택 사항*, 기본값 `False`) : `False`인 경우, 이 함수는 최종 video processor 객체만 반환합니다. `True`인 경우, 이 함수는 `Tuple(video_processor, unused_kwargs)`를 반환하며, 여기서 *unused_kwargs*는 video processor 속성이 아닌 키/값 쌍으로 구성된 사전입니다. 즉, `kwargs` 중 `video_processor`를 업데이트하는 데 사용되지 않고 무시되는 부분입니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

kwargs (`dict[str, Any]`, *선택*) : kwargs의 값 중 video processor 속성에 해당하는 키는 로드된 값을 덮어쓰는 데 사용됩니다. video processor 속성이 *아닌* 키/값 쌍에 대한 동작은 `return_unused_kwargs` 키워드 매개변수로 제어됩니다.

pretrained model 어휘에서 라이브러리의 비디오 프로세서 클래스 중 하나를 인스턴스화합니다.

인스턴스화할 video processor 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 값)에 따라 선택되며, 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백됩니다:

- **cohere_compass** -- `{'torchvision': 'CohereCompassVideoProcessor'}` (CohereCompassConfig 모델)
- **cosmos3_edge** -- `{'torchvision': 'Cosmos3EdgeVideoProcessor'}` (Cosmos3EdgeConfig 모델)
- **cosmos3_omni** -- `{'torchvision': 'Qwen3VLVideoProcessor'}` (Cosmos3OmniConfig 모델)
- **dots3_note** -- `{'torchvision': 'Qwen3VLVideoProcessor'}` (Dots3NoteConfig 모델)
- **embedding_gemma2** -- `{'torchvision': 'EmbeddingGemma2VideoProcessor'}` (EmbeddingGemma2Config 모델)
- **ernie4_5_vl_moe** -- `{'torchvision': 'Ernie4_5_VLMoeVideoProcessor'}` (Ernie4_5_VLMoeConfig 모델)
- **exaone4_5** -- `{'torchvision': 'Qwen2VLVideoProcessor'}` (Exaone4_5_Config 모델)
- **gemma4** -- `{'torchvision': 'Gemma4VideoProcessor'}` (Gemma4Config 모델)
- **gemma4_unified** -- `{'torchvision': 'Gemma4UnifiedVideoProcessor'}` (Gemma4UnifiedConfig 모델)
- **glm46v** -- `{'torchvision': 'Glm46VVideoProcessor'}` (Glm46VConfig 모델)
- **glm4v** -- `{'torchvision': 'Glm4vVideoProcessor'}` (Glm4vConfig 모델)
- **glm5_next** -- `{'torchvision': 'Glm5NextVideoProcessor'}` (Glm5NextConfig 모델)
- **glmga** -- `{'torchvision': 'GlmgaVideoProcessor'}` (GlmgaConfig 모델)
- **hyperclovax_vision_v2** -- `{'torchvision': 'Qwen2VLVideoProcessor'}` (HyperCLOVAXVisionV2Config 모델)
- **instructblip** -- `{'torchvision': 'InstructBlipVideoVideoProcessor'}` (InstructBlipConfig 모델)
- **instructblipvideo** -- `{'torchvision': 'InstructBlipVideoVideoProcessor'}` (InstructBlipVideoConfig 모델)
- **internvl** -- `{'torchvision': 'InternVLVideoProcessor'}` (InternVLConfig 모델)
- **kimi_k25** -- `{'torchvision': 'Kimi_K25VideoProcessor'}` (Kimi_K25Config 모델)
- **lfm3_vl** -- `{'torchvision': 'Qwen3VLVideoProcessor'}` (Lfm3VlConfig 모델)
- **llava_next_video** -- `{'torchvision': 'LlavaNextVideoVideoProcessor'}` (LlavaNextVideoConfig 모델)
- **llava_onevision** -- `{'torchvision': 'LlavaOnevisionVideoProcessor'}` (LlavaOnevisionConfig 모델)
- **minicpm_ocr** -- `{'torchvision': 'MiniCPMOcrVideoProcessor'}` (MiniCPMOcrConfig 모델)
- **minicpmv4_6** -- `{'torchvision': 'MiniCPMV4_6VideoProcessor'}` (MiniCPMV4_6Config 모델)
- **minicpmv4_7** -- `{'torchvision': 'MiniCPMV4_6VideoProcessor'}` (MiniCPMV4_7Config 모델)
- **minimax_m3_vl** -- `{'torchvision': 'MiniMaxM3VLVideoProcessor'}` (MiniMaxM3VLConfig 모델)
- **molmo2** -- `{'torchvision': 'Molmo2VideoProcessor'}` (Molmo2Config 모델)
- **muse_glimmer** -- `{'torchvision': 'MuseGlimmerVideoProcessor'}` (MuseGlimmerConfig 모델)
- **muse_spark** -- `{'torchvision': 'MuseGlimmerVideoProcessor'}` (MuseSparkConfig 모델)
- **nemotron_h_omni** -- `{'torchvision': 'NemotronH_Omni_Reasoning_V3VideoProcessor'}` (NemotronH_Omni_Reasoning_V3_Config 모델)
- **pe_audio_video** -- `{'torchvision': 'PeVideoVideoProcessor'}` (PeAudioVideoConfig 모델)
- **pe_video** -- `{'torchvision': 'PeVideoVideoProcessor'}` (PeVideoConfig 모델)
- **perception_lm** -- `{'torchvision': 'PerceptionLMVideoProcessor'}` (PerceptionLMConfig 모델)
- **qwen2_5_omni** -- `{'torchvision': 'Qwen2VLVideoProcessor'}` (Qwen2_5OmniConfig 모델)
- **qwen2_5_vl** -- `{'torchvision': 'Qwen2VLVideoProcessor'}` (Qwen2_5_VLConfig 모델)
- **qwen2_vl** -- `{'torchvision': 'Qwen2VLVideoProcessor'}` (Qwen2VLConfig 모델)
- **qwen3_5** -- `{'torchvision': 'Qwen3VLVideoProcessor'}` (Qwen3_5Config 모델)
- **qwen3_5_moe** -- `{'torchvision': 'Qwen3VLVideoProcessor'}` (Qwen3_5MoeConfig 모델)
- **qwen3_omni_moe** -- `{'torchvision': 'Qwen2VLVideoProcessor'}` (Qwen3OmniMoeConfig 모델)
- **qwen3_vl** -- `{'torchvision': 'Qwen3VLVideoProcessor'}` (Qwen3VLConfig 모델)
- **qwen3_vl_moe** -- `{'torchvision': 'Qwen3VLVideoProcessor'}` (Qwen3VLMoeConfig 모델)
- **qwen4_exp** -- `{'torchvision': 'Qwen3VLVideoProcessor'}` (Qwen4ExpConfig 모델)
- **sam2_video** -- `{'torchvision': 'Sam2VideoVideoProcessor'}` (Sam2VideoConfig 모델)
- **smolvlm** -- `{'torchvision': 'SmolVLMVideoProcessor'}` (SmolVLMConfig 모델)
- **timesformer** -- `{'torchvision': 'TimesformerVideoProcessor'}` (TimesformerConfig 모델)
- **video_llama_3** -- `{'torchvision': 'VideoLlama3VideoProcessor'}` (VideoLlama3Config 모델)
- **video_llava** -- `{'torchvision': 'VideoLlavaVideoProcessor'}` (VideoLlavaConfig 모델)
- **videomae** -- `{'torchvision': 'VideoMAEVideoProcessor'}` (VideoMAEConfig 모델)
- **videomt** -- `{'torchvision': 'VideomtVideoProcessor'}` (VideomtConfig 모델)
- **videoprism** -- `{'torchvision': 'LlavaOnevisionVideoProcessor'}` (VideoPrismConfig 모델)
- **vivit** -- `{'torchvision': 'VivitVideoProcessor'}` (VivitConfig 모델)
- **vjepa2** -- `{'torchvision': 'VJEPA2VideoProcessor'}` (VJEPA2Config 모델)

비공개 모델을 사용하려면 `token=True`를 전달해야 합니다.

예:

```python
>>> from transformers import AutoVideoProcessor

>>> # Download video processor from huggingface.co and cache.
>>> video_processor = AutoVideoProcessor.from_pretrained("llava-hf/llava-onevision-qwen2-0.5b-ov-hf")

>>> # If video processor files are in a directory (e.g. video processor was saved using *save_pretrained('./test/saved_model/')*)
>>> # video_processor = AutoVideoProcessor.from_pretrained("./test/saved_model/")
```

#### register[[transformers.AutoVideoProcessor.register]]

```python
register(config_class, video_processor_class: type | None = None, video_processor_classes: dict[str, type] | None = None, exist_ok: bool = False)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/video_processing_auto.py#L466)

**매개변수:**

config_class ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 등록할 모델에 해당하는 설정입니다.

video_processor_class ([BaseVideoProcessor](/docs/transformers/v5.23.0/ko/main_classes/video_processor#transformers.BaseVideoProcessor)) : 등록할 video processor입니다.

video_processor_classes (`dict[str, type]`, *선택*) : 백엔드 이름을 비디오 프로세서 클래스로 매핑하는 사전입니다. 사용자 정의 백엔드를 등록할 수 있습니다. 예: `{"pil": MyPilVideoProcessor, "torchvision": MyTorchvisionVideoProcessor, "custom": MyCustomVideoProcessor}`

exist_ok (`bool`, *선택 사항*, 기본값 `False`) : `True`인 경우 기존 등록을 덮어쓰는 것을 허용합니다.

이 클래스에 새로운 video processor를 등록합니다.

## AutoProcessor[[transformers.AutoProcessor]]

#### transformers.AutoProcessor[[transformers.AutoProcessor]]

```python
transformers.AutoProcessor()
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/processing_auto.py#L127)

이것은 [AutoProcessor.from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoProcessor.from_pretrained) 클래스 메서드로 생성될 때 라이브러리의 프로세서 클래스 중 하나로 인스턴스화되는 범용 프로세서 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_pretrained[[transformers.AutoProcessor.from_pretrained]]

```python
from_pretrained(pretrained_model_name_or_path, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/processing_auto.py#L141)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습된 feature_extractor의 *model id*인 문자열. - `save_pretrained()` 메서드로 저장된 processor 파일이 포함된 *디렉터리*의 경로, 예: `./my_model_directory/`.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 feature extractor를 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : feature extractor 파일을 (재)다운로드를 강제하고, 캐시된 버전이 존재하는 경우 이를 덮어쓸지 여부.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}.` 각 요청에 프록시가 사용됩니다.

token (`str` 또는 *bool*, *선택*) : 원격 파일에 대한 HTTP bearer 인증으로 사용할 토큰입니다. `True`인 경우, `hf auth login` 실행 시 생성된 토큰(`~/.huggingface`에 저장됨)을 사용합니다.

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

return_unused_kwargs (`bool`, *선택 사항*, 기본값 `False`) : `False`인 경우, 이 함수는 최종 feature extractor 객체만 반환합니다. `True`인 경우, 이 함수는 `Tuple(feature_extractor, unused_kwargs)`를 반환하며, 여기서 *unused_kwargs*는 feature extractor 속성이 아닌 키/값 쌍으로 구성된 사전입니다. 즉, `kwargs` 중 `feature_extractor`를 업데이트하는 데 사용되지 않고 무시되는 부분입니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

kwargs (`dict[str, Any]`, *선택*) : kwargs의 값 중 feature extractor 속성에 해당하는 키는 로드된 값을 덮어쓰는 데 사용됩니다. feature extractor 속성이 *아닌* 키/값 쌍에 대한 동작은 `return_unused_kwargs` 키워드 매개변수로 제어됩니다.

…

인스턴스화할 프로세서 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된)에 따라 선택됩니다:

- **aimv2** -- [CLIPProcessor](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPProcessor) (Aimv2Config 모델)
- **align** -- [AlignProcessor](/docs/transformers/v5.23.0/ko/model_doc/align#transformers.AlignProcessor) (AlignConfig 모델)
- **altclip** -- [AltCLIPProcessor](/docs/transformers/v5.23.0/ko/model_doc/altclip#transformers.AltCLIPProcessor) (AltCLIPConfig 모델)
- **aria** -- [AriaProcessor](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaProcessor) (AriaConfig 모델)
- **audioflamingo3** -- [AudioFlamingo3Processor](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3Processor) (AudioFlamingo3Config 모델)
- **aya_vision** -- [AyaVisionProcessor](/docs/transformers/v5.23.0/ko/model_doc/aya_vision#transformers.AyaVisionProcessor) (AyaVisionConfig 모델)
- **bark** -- [BarkProcessor](/docs/transformers/v5.23.0/ko/model_doc/bark#transformers.BarkProcessor) (BarkConfig 모델)
- **blip** -- [BlipProcessor](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipProcessor) (BlipConfig 모델)
- **blip-2** -- [Blip2Processor](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2Processor) (Blip2Config 모델)
- **bridgetower** -- [BridgeTowerProcessor](/docs/transformers/v5.23.0/ko/model_doc/bridgetower#transformers.BridgeTowerProcessor) (BridgeTowerConfig 모델)
- **bros** -- [BrosProcessor](/docs/transformers/v5.23.0/ko/model_doc/bros#transformers.BrosProcessor) (BrosConfig 모델)
- **canary** -- [CanaryProcessor](/docs/transformers/v5.23.0/ko/model_doc/canary#transformers.CanaryProcessor) (CanaryConfig 모델)
- **chameleon** -- [ChameleonProcessor](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonProcessor) (ChameleonConfig 모델)
- **chinese_clip** -- [ChineseCLIPProcessor](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPProcessor) (ChineseCLIPConfig 모델)
- **clap** -- [ClapProcessor](/docs/transformers/v5.23.0/ko/model_doc/clap#transformers.ClapProcessor) (ClapConfig 모델)
- **clip** -- [CLIPProcessor](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPProcessor) (CLIPConfig 모델)
- **clipseg** -- [CLIPSegProcessor](/docs/transformers/v5.23.0/ko/model_doc/clipseg#transformers.CLIPSegProcessor) (CLIPSegConfig 모델)
- **clvp** -- [ClvpProcessor](/docs/transformers/v5.23.0/ko/model_doc/clvp#transformers.ClvpProcessor) (ClvpConfig 모델)
- **cohere2_vision** -- [Cohere2VisionProcessor](/docs/transformers/v5.23.0/ko/model_doc/cohere2_vision#transformers.Cohere2VisionProcessor) (Cohere2VisionConfig 모델)
- **cohere_asr** -- [CohereAsrProcessor](/docs/transformers/v5.23.0/ko/model_doc/cohere_asr#transformers.CohereAsrProcessor) (CohereAsrConfig 모델)
- **cohere_compass** -- [CohereCompassProcessor](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassProcessor) (CohereCompassConfig 모델)
- **colmodernvbert** -- [ColModernVBertProcessor](/docs/transformers/v5.23.0/ko/model_doc/colmodernvbert#transformers.ColModernVBertProcessor) (ColModernVBertConfig 모델)
- **colpali** -- [ColPaliProcessor](/docs/transformers/v5.23.0/ko/model_doc/colpali#transformers.ColPaliProcessor) (ColPaliConfig 모델)
- **colqwen2** -- [ColQwen2Processor](/docs/transformers/v5.23.0/ko/model_doc/colqwen2#transformers.ColQwen2Processor) (ColQwen2Config 모델)
- **cosmos3_edge** -- [Cosmos3EdgeProcessor](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeProcessor) (Cosmos3EdgeConfig 모델)
- **csm** -- [CsmProcessor](/docs/transformers/v5.23.0/ko/model_doc/csm#transformers.CsmProcessor) (CsmConfig 모델)
- **deepseek_ocr2** -- [DeepseekOcr2Processor](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2Processor) (DeepseekOcr2Config 모델)
- **deepseek_vl** -- [DeepseekVLProcessor](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl#transformers.DeepseekVLProcessor) (DeepseekVLConfig 모델)
- **deepseek_vl_hybrid** -- [DeepseekVLHybridProcessor](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl_hybrid#transformers.DeepseekVLHybridProcessor) (DeepseekVLHybridConfig 모델)
- **dia** -- [DiaProcessor](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaProcessor) (DiaConfig 모델)
- **diffusion_gemma** -- [Gemma4Processor](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4Processor) (DiffusionGemmaConfig 모델)
- **dots3_note** -- [Dots3NoteProcessor](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteProcessor) (Dots3NoteConfig 모델)
- **edgetam** -- [Sam2Processor](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2Processor) (EdgeTamConfig 모델)
- **embedding_gemma2** -- [EmbeddingGemma2Processor](/docs/transformers/v5.23.0/ko/model_doc/embedding_gemma2#transformers.EmbeddingGemma2Processor) (EmbeddingGemma2Config 모델)
- **emu3** -- [Emu3Processor](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3Processor) (Emu3Config 모델)
- **ernie4_5_vl_moe** -- [Ernie4_5_VLMoeProcessor](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeProcessor) (Ernie4_5_VLMoeConfig 모델)
- **evolla** -- [EvollaProcessor](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaProcessor) (EvollaConfig 모델)
- **exaone4_5** -- [Exaone4_5_Processor](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_Processor) (Exaone4_5_Config 모델)
- **flava** -- [FlavaProcessor](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaProcessor) (FlavaConfig 모델)
- **florence2** -- [Florence2Processor](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2Processor) (Florence2Config 모델)
- **fun_asr_nano** -- [FunAsrNanoProcessor](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoProcessor) (FunAsrNanoConfig 모델)
- **fuyu** -- [FuyuProcessor](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuProcessor) (FuyuConfig 모델)
- **gemma3** -- [Gemma3Processor](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Processor) (Gemma3Config 모델)
- **gemma3n** -- [Gemma3nProcessor](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nProcessor) (Gemma3nConfig 모델)
- **gemma4** -- [Gemma4Processor](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4Processor) (Gemma4Config 모델)
- **gemma4_unified** -- [Gemma4UnifiedProcessor](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedProcessor) (Gemma4UnifiedConfig 모델)
- **git** -- [GitProcessor](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitProcessor) (GitConfig 모델)
- **glm46v** -- [Glm46VProcessor](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VProcessor) (Glm46VConfig 모델)
- **glm4v** -- [Glm4vProcessor](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vProcessor) (Glm4vConfig 모델)
- **glm4v_moe** -- [Glm4vProcessor](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vProcessor) (Glm4vMoeConfig 모델)
- **glm5_next** -- [Glm5NextProcessor](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextProcessor) (Glm5NextConfig 모델)
- **glm_image** -- [GlmImageProcessor](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageProcessor) (GlmImageConfig 모델)
- **glmasr** -- [GlmAsrProcessor](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrProcessor) (GlmAsrConfig 모델)
- **got_ocr2** -- [GotOcr2Processor](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2Processor) (GotOcr2Config 모델)
- **gradium_phonon_tts** -- [PocketTTSProcessor](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts#transformers.PocketTTSProcessor) (GradiumPhononTTSConfig 모델)
- **granite4_vision** -- [Granite4VisionProcessor](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionProcessor) (Granite4VisionConfig 모델)
- **granite_docling** -- [GraniteDoclingProcessor](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingProcessor) (GraniteDoclingConfig 모델)
- **granite_speech** -- [GraniteSpeechProcessor](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechProcessor) (GraniteSpeechConfig 모델)
- **granite_speech5_ctc** -- [GraniteSpeech5Processor](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5Processor) (GraniteSpeech5CTCConfig 모델)
- **granite_speech_plus** -- [GraniteSpeechProcessor](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechProcessor) (GraniteSpeechPlusConfig 모델)
- **grounding-dino** -- [GroundingDinoProcessor](/docs/transformers/v5.23.0/ko/model_doc/grounding-dino#transformers.GroundingDinoProcessor) (GroundingDinoConfig 모델)
- **groupvit** -- [CLIPProcessor](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPProcessor) (GroupViTConfig 모델)
- **higgs_audio_v2** -- [HiggsAudioV2Processor](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2#transformers.HiggsAudioV2Processor) (HiggsAudioV2Config 모델)
- **hubert** -- [Wav2Vec2Processor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Processor) (HubertConfig 모델)
- **hunyuan_vl** -- [HunYuanVLProcessor](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLProcessor) (HunYuanVLConfig 모델)
- **hyperclovax_vision_v2** -- [HyperCLOVAXVisionV2Processor](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax_vision_v2#transformers.HyperCLOVAXVisionV2Processor) (HyperCLOVAXVisionV2Config 모델)
- **idefics** -- [IdeficsProcessor](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsProcessor) (IdeficsConfig 모델)
- **idefics2** -- [Idefics2Processor](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2Processor) (Idefics2Config 모델)
- **idefics3** -- [Idefics3Processor](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3Processor) (Idefics3Config 모델)
- **inkling2** -- [Inkling2Processor](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2Processor) (Inkling2Config 모델)
- **inkling_mm_model** -- [InklingProcessor](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingProcessor) (InklingConfig 모델)
- **instructblip** -- [InstructBlipProcessor](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipProcessor) (InstructBlipConfig 모델)
- **instructblipvideo** -- [InstructBlipVideoProcessor](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoProcessor) (InstructBlipVideoConfig 모델)
- **internvl** -- [InternVLProcessor](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLProcessor) (InternVLConfig 모델)
- **janus** -- [JanusProcessor](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusProcessor) (JanusConfig 모델)
- **kimi_k25** -- [Kimi_K25Processor](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25Processor) (Kimi_K25Config 모델)
- **kyutai_speech_to_text** -- [KyutaiSpeechToTextProcessor](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextProcessor) (KyutaiSpeechToTextConfig 모델)
- **lasr_ctc** -- [LasrProcessor](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrProcessor) (LasrCTCConfig 모델)
- **lasr_encoder** -- [LasrProcessor](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrProcessor) (LasrEncoderConfig 모델)
- **layoutlmv2** -- [LayoutLMv2Processor](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2Processor) (LayoutLMv2Config 모델)
- **layoutlmv3** -- [LayoutLMv3Processor](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3Processor) (LayoutLMv3Config 모델)
- **layoutxlm** -- [LayoutXLMProcessor](/docs/transformers/v5.23.0/ko/model_doc/layoutxlm#transformers.LayoutXLMProcessor) (LayoutXLMConfig 모델)
- **lfm2_vl** -- [Lfm2VlProcessor](/docs/transformers/v5.23.0/ko/model_doc/lfm2_vl#transformers.Lfm2VlProcessor) (Lfm2VlConfig 모델)
- **lfm3_vl** -- [Lfm3VlProcessor](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlProcessor) (Lfm3VlConfig 모델)
- **lighton_ocr** -- [LightOnOcrProcessor](/docs/transformers/v5.23.0/ko/model_doc/lighton_ocr#transformers.LightOnOcrProcessor) (LightOnOcrConfig 모델)
- **llama4** -- [Llama4Processor](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4Processor) (Llama4Config 모델)
- **llava** -- [LlavaProcessor](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaProcessor) (LlavaConfig 모델)
- **llava_next** -- [LlavaNextProcessor](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextProcessor) (LlavaNextConfig 모델)
- **llava_next_video** -- [LlavaNextVideoProcessor](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoProcessor) (LlavaNextVideoConfig 모델)
- **llava_onevision** -- [LlavaOnevisionProcessor](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionProcessor) (LlavaOnevisionConfig 모델)
- **markuplm** -- [MarkupLMProcessor](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMProcessor) (MarkupLMConfig 모델)
- **metaclip_2** -- [CLIPProcessor](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPProcessor) (MetaClip2Config 모델)
- **mgp-str** -- [MgpstrProcessor](/docs/transformers/v5.23.0/ko/model_doc/mgp-str#transformers.MgpstrProcessor) (MgpstrConfig 모델)
- **minicpm_ocr** -- [MiniCPMOcrProcessor](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrProcessor) (MiniCPMOcrConfig 모델)
- **minicpmv4_6** -- [MiniCPMV4_6Processor](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6Processor) (MiniCPMV4_6Config 모델)
- **minicpmv4_7** -- [MiniCPMV4_7Processor](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7Processor) (MiniCPMV4_7Config 모델)
- **minimax_m3_vl** -- [MiniMaxM3VLProcessor](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLProcessor) (MiniMaxM3VLConfig 모델)
- **mistral3** -- [PixtralProcessor](/docs/transformers/v5.23.0/ko/model_doc/pixtral#transformers.PixtralProcessor) (Mistral3Config 모델)
- **mistral_large4** -- [PixtralProcessor](/docs/transformers/v5.23.0/ko/model_doc/pixtral#transformers.PixtralProcessor) (MistralLarge4Config 모델)
- **mllama** -- [MllamaProcessor](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaProcessor) (MllamaConfig 모델)
- **mm-grounding-dino** -- [GroundingDinoProcessor](/docs/transformers/v5.23.0/ko/model_doc/grounding-dino#transformers.GroundingDinoProcessor) (MMGroundingDinoConfig 모델)
- **modernvbert** -- [Idefics3Processor](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3Processor) (ModernVBertConfig 모델)
- **molmo2** -- [Molmo2Processor](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2Processor) (Molmo2Config 모델)
- **moonshine** -- [Wav2Vec2Processor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Processor) (MoonshineConfig 모델)
- **moonshine_streaming** -- [MoonshineStreamingProcessor](/docs/transformers/v5.23.0/ko/model_doc/moonshine_streaming#transformers.MoonshineStreamingProcessor) (MoonshineStreamingConfig 모델)
- **moss_transcribe_diarize** -- [MossTranscribeDiarizeProcessor](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeProcessor) (MossTranscribeDiarizeConfig 모델)
- **muse_glimmer** -- [MuseGlimmerProcessor](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerProcessor) (MuseGlimmerConfig 모델)
- **muse_spark** -- [MuseSparkProcessor](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkProcessor) (MuseSparkConfig 모델)
- **musicflamingo** -- [MusicFlamingoProcessor](/docs/transformers/v5.23.0/ko/model_doc/musicflamingo#transformers.MusicFlamingoProcessor) (MusicFlamingoConfig 모델)
- **musicgen** -- [MusicgenProcessor](/docs/transformers/v5.23.0/ko/model_doc/musicgen#transformers.MusicgenProcessor) (MusicgenConfig 모델)
- **musicgen_melody** -- [MusicgenMelodyProcessor](/docs/transformers/v5.23.0/ko/model_doc/musicgen_melody#transformers.MusicgenMelodyProcessor) (MusicgenMelodyConfig 모델)
- **nemo_speechlm** -- [Nemotron3_5TranscribeProcessor](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeProcessor) (Nemotron3_5TranscribeConfig 모델)
- **nemotron3_5_asr** -- [Nemotron3_5AsrProcessor](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_asr#transformers.Nemotron3_5AsrProcessor) (Nemotron3_5AsrConfig 모델)
- **nemotron3_diarization** -- [Nemotron3DiarizationProcessor](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationProcessor) (Nemotron3DiarizationConfig 모델)
- **nemotron_asr_streaming** -- [NemotronAsrStreamingProcessor](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingProcessor) (NemotronAsrStreamingConfig 모델)
- **nemotron_h_omni** -- [NemotronH_Omni_Reasoning_V3Processor](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h_omni#transformers.NemotronH_Omni_Reasoning_V3Processor) (NemotronH_Omni_Reasoning_V3_Config 모델)
- **neomme** -- [NeoMMEProcessor](/docs/transformers/v5.23.0/ko/model_doc/neomme#transformers.NeoMMEProcessor) (NeoMMEConfig 모델)
- **nougat** -- [NougatProcessor](/docs/transformers/v5.23.0/ko/model_doc/nougat#transformers.NougatProcessor) (NougatConfig 모델)
- **omdet-turbo** -- [OmDetTurboProcessor](/docs/transformers/v5.23.0/ko/model_doc/omdet-turbo#transformers.OmDetTurboProcessor) (OmDetTurboConfig 모델)
- **oneformer** -- [OneFormerProcessor](/docs/transformers/v5.23.0/ko/model_doc/oneformer#transformers.OneFormerProcessor) (OneFormerConfig 모델)
- **ovis2** -- [Ovis2Processor](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2Processor) (Ovis2Config 모델)
- **owlv2** -- [Owlv2Processor](/docs/transformers/v5.23.0/ko/model_doc/owlv2#transformers.Owlv2Processor) (Owlv2Config 모델)
- **owlvit** -- [OwlViTProcessor](/docs/transformers/v5.23.0/ko/model_doc/owlvit#transformers.OwlViTProcessor) (OwlViTConfig 모델)
- **paddleocr_vl** -- [PaddleOCRVLProcessor](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVLProcessor) (PaddleOCRVLConfig 모델)
- **paligemma** -- [PaliGemmaProcessor](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaProcessor) (PaliGemmaConfig 모델)
- **parakeet_ctc** -- [ParakeetProcessor](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetProcessor) (ParakeetCTCConfig 모델)
- **parakeet_rnnt** -- [ParakeetProcessor](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetProcessor) (ParakeetRNNTConfig 모델)
- **parakeet_tdt** -- [ParakeetProcessor](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetProcessor) (ParakeetTDTConfig 모델)
- **pe_audio** -- [PeAudioProcessor](/docs/transformers/v5.23.0/ko/model_doc/pe_audio#transformers.PeAudioProcessor) (PeAudioConfig 모델)
- **pe_audio_video** -- [PeAudioVideoProcessor](/docs/transformers/v5.23.0/ko/model_doc/pe_audio_video#transformers.PeAudioVideoProcessor) (PeAudioVideoConfig 모델)
- **pe_video** -- [PeVideoProcessor](/docs/transformers/v5.23.0/ko/model_doc/pe_video#transformers.PeVideoProcessor) (PeVideoConfig 모델)
- **perception_lm** -- [PerceptionLMProcessor](/docs/transformers/v5.23.0/ko/model_doc/perception_lm#transformers.PerceptionLMProcessor) (PerceptionLMConfig 모델)
- **phi4_multimodal** -- [Phi4MultimodalProcessor](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalProcessor) (Phi4MultimodalConfig 모델)
- **pi0** -- [PI0Processor](/docs/transformers/v5.23.0/ko/model_doc/pi0#transformers.PI0Processor) (PI0Config 모델)
- **pix2struct** -- [Pix2StructProcessor](/docs/transformers/v5.23.0/ko/model_doc/pix2struct#transformers.Pix2StructProcessor) (Pix2StructConfig 모델)
- **pixtral** -- [PixtralProcessor](/docs/transformers/v5.23.0/ko/model_doc/pixtral#transformers.PixtralProcessor) (PixtralVisionConfig 모델)
- **pocket_tts** -- [PocketTTSProcessor](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts#transformers.PocketTTSProcessor) (PocketTTSConfig 모델)
- **pop2piano** -- [Pop2PianoProcessor](/docs/transformers/v5.23.0/ko/model_doc/pop2piano#transformers.Pop2PianoProcessor) (Pop2PianoConfig 모델)
- **pp_chart2table** -- [PPChart2TableProcessor](/docs/transformers/v5.23.0/ko/model_doc/pp_chart2table#transformers.PPChart2TableProcessor) (PPChart2TableConfig 모델)
- **pp_formulanet** -- [PPFormulaNetProcessor](/docs/transformers/v5.23.0/ko/model_doc/pp_formulanet#transformers.PPFormulaNetProcessor) (PPFormulaNetConfig 모델)
- **qianfan_ocr** -- [QianfanOCRProcessor](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRProcessor) (QianfanOCRConfig 모델)
- **qwen2_5_omni** -- [Qwen2_5OmniProcessor](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniProcessor) (Qwen2_5OmniConfig 모델)
- **qwen2_5_vl** -- [Qwen2_5_VLProcessor](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLProcessor) (Qwen2_5_VLConfig 모델)
- **qwen2_audio** -- [Qwen2AudioProcessor](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioProcessor) (Qwen2AudioConfig 모델)
- **qwen2_vl** -- [Qwen2VLProcessor](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLProcessor) (Qwen2VLConfig 모델)
- **qwen3_5** -- [Qwen3VLProcessor](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLProcessor) (Qwen3_5Config 모델)
- **qwen3_5_moe** -- [Qwen3VLProcessor](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLProcessor) (Qwen3_5MoeConfig 모델)
- **qwen3_asr** -- [Qwen3ASRProcessor](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRProcessor) (Qwen3ASRConfig 모델)
- **qwen3_omni_moe** -- [Qwen3OmniMoeProcessor](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeProcessor) (Qwen3OmniMoeConfig 모델)
- **qwen3_vl** -- [Qwen3VLProcessor](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLProcessor) (Qwen3VLConfig 모델)
- **qwen3_vl_moe** -- [Qwen3VLProcessor](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLProcessor) (Qwen3VLMoeConfig 모델)
- **qwen4_exp** -- [Qwen3VLProcessor](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLProcessor) (Qwen4ExpConfig 모델)
- **sam** -- [SamProcessor](/docs/transformers/v5.23.0/ko/model_doc/sam#transformers.SamProcessor) (SamConfig 모델)
- **sam2** -- [Sam2Processor](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2Processor) (Sam2Config 모델)
- **sam2_video** -- [Sam2VideoProcessor](/docs/transformers/v5.23.0/ko/model_doc/sam2_video#transformers.Sam2VideoProcessor) (Sam2VideoConfig 모델)
- **sam3** -- [Sam3Processor](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3Processor) (Sam3Config 모델)
- **sam3_lite_text** -- [Sam3Processor](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3Processor) (Sam3LiteTextConfig 모델)
- **sam3_tracker** -- [Sam3TrackerProcessor](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker#transformers.Sam3TrackerProcessor) (Sam3TrackerConfig 모델)
- **sam3_tracker_video** -- [Sam3TrackerVideoProcessor](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker_video#transformers.Sam3TrackerVideoProcessor) (Sam3TrackerVideoConfig 모델)
- **sam3_video** -- [Sam3VideoProcessor](/docs/transformers/v5.23.0/ko/model_doc/sam3_video#transformers.Sam3VideoProcessor) (Sam3VideoConfig 모델)
- **sam_hq** -- [SamHQProcessor](/docs/transformers/v5.23.0/ko/model_doc/sam_hq#transformers.SamHQProcessor) (SamHQConfig 모델)
- **seamless_m4t** -- [SeamlessM4TProcessor](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TProcessor) (SeamlessM4TConfig 모델)
- **sew** -- [Wav2Vec2Processor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Processor) (SEWConfig 모델)
- **sew-d** -- [Wav2Vec2Processor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Processor) (SEWDConfig 모델)
- **shieldgemma2** -- [ShieldGemma2Processor](/docs/transformers/v5.23.0/ko/model_doc/shieldgemma2#transformers.ShieldGemma2Processor) (ShieldGemma2Config 모델)
- **siglip** -- [SiglipProcessor](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipProcessor) (SiglipConfig 모델)
- **siglip2** -- [Siglip2Processor](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2Processor) (Siglip2Config 모델)
- **smolvlm** -- [SmolVLMProcessor](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMProcessor) (SmolVLMConfig 모델)
- **speech_to_text** -- [Speech2TextProcessor](/docs/transformers/v5.23.0/ko/model_doc/speech_to_text#transformers.Speech2TextProcessor) (Speech2TextConfig 모델)
- **speecht5** -- [SpeechT5Processor](/docs/transformers/v5.23.0/ko/model_doc/speecht5#transformers.SpeechT5Processor) (SpeechT5Config 모델)
- **step3p7** -- [Step3p7Processor](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7Processor) (Step3p7Config 모델)
- **t5gemma2** -- [Gemma3Processor](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Processor) (T5Gemma2Config 모델)
- **t5gemma2_encoder** -- [Gemma3Processor](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Processor) (T5Gemma2EncoderConfig 모델)
- **tipsv2** -- [Tipsv2Processor](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2Processor) (Tipsv2Config 모델)
- **trocr** -- [TrOCRProcessor](/docs/transformers/v5.23.0/ko/model_doc/trocr#transformers.TrOCRProcessor) (TrOCRConfig 모델)
- **tvp** -- [TvpProcessor](/docs/transformers/v5.23.0/ko/model_doc/tvp#transformers.TvpProcessor) (TvpConfig 모델)
- **udop** -- [UdopProcessor](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopProcessor) (UdopConfig 모델)
- **ultravbert** -- [Idefics3Processor](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3Processor) (UltraVBertConfig 모델)
- **unispeech** -- [Wav2Vec2Processor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Processor) (UniSpeechConfig 모델)
- **unispeech-sat** -- [Wav2Vec2Processor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Processor) (UniSpeechSatConfig 모델)
- **unlimited_ocr** -- [UnlimitedOcrProcessor](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrProcessor) (UnlimitedOcrConfig 모델)
- **vibevoice** -- [VibeVoiceProcessor](/docs/transformers/v5.23.0/ko/model_doc/vibevoice#transformers.VibeVoiceProcessor) (VibeVoiceConfig 모델)
- **vibevoice_asr** -- [VibeVoiceAsrProcessor](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrProcessor) (VibeVoiceAsrConfig 모델)
- **video_llama_3** -- [VideoLlama3Processor](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3Processor) (VideoLlama3Config 모델)
- **video_llava** -- [VideoLlavaProcessor](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaProcessor) (VideoLlavaConfig 모델)
- **videoprism** -- [VideoPrismProcessor](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismProcessor) (VideoPrismConfig 모델)
- **vilt** -- [ViltProcessor](/docs/transformers/v5.23.0/ko/model_doc/vilt#transformers.ViltProcessor) (ViltConfig 모델)
- **vipllava** -- [LlavaProcessor](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaProcessor) (VipLlavaConfig 모델)
- **vision-text-dual-encoder** -- [VisionTextDualEncoderProcessor](/docs/transformers/v5.23.0/ko/model_doc/vision-text-dual-encoder#transformers.VisionTextDualEncoderProcessor) (VisionTextDualEncoderConfig 모델)
- **voxtral** -- [VoxtralProcessor](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralProcessor) (VoxtralConfig 모델)
- **voxtral_realtime** -- [VoxtralRealtimeProcessor](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeProcessor) (VoxtralRealtimeConfig 모델)
- **wav2vec2** -- [Wav2Vec2Processor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Processor) (Wav2Vec2Config 모델)
- **wav2vec2-bert** -- [Wav2Vec2BertProcessor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertProcessor) (Wav2Vec2BertConfig 모델)
- **wav2vec2-conformer** -- [Wav2Vec2Processor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Processor) (Wav2Vec2ConformerConfig 모델)
- **wavlm** -- [Wav2Vec2Processor](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Processor) (WavLMConfig 모델)
- **whisper** -- [WhisperProcessor](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperProcessor) (WhisperConfig 모델)

비공개 모델을 사용하려면 `token=True`를 전달해야 합니다.

예:

```python
>>> from transformers import AutoProcessor

>>> # Download processor from huggingface.co and cache.
>>> processor = AutoProcessor.from_pretrained("facebook/wav2vec2-base-960h")

>>> # If processor files are in a directory (e.g. processor was saved using *save_pretrained('./test/saved_model/')*)
>>> # processor = AutoProcessor.from_pretrained("./test/saved_model/")
```

#### register[[transformers.AutoProcessor.register]]

```python
register(config_class, processor_class, exist_ok = False)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/processing_auto.py#L373)

**매개변수:**

config_class ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 등록할 모델에 해당하는 설정입니다.

processor_class ([ProcessorMixin](/docs/transformers/v5.23.0/ko/main_classes/processors#transformers.ProcessorMixin)) : 등록할 프로세서입니다.

이 클래스에 새 프로세서를 등록합니다.

## 범용 모델 클래스

auto 클래스는 특정 헤드가 없는 기본 모델 클래스를 인스턴스화하는 데 사용할 수 있습니다.

### AutoModel[[transformers.AutoModel]]

#### transformers.AutoModel[[transformers.AutoModel]]

```python
transformers.AutoModel(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2333)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때 라이브러리의 기본 모델 클래스 중 하나로 인스턴스화되는 범용 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModel.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : The model class to instantiate is selected based on the configuration class:  - [ASTConfig](/docs/transformers/v5.23.0/ko/model_doc/audio-spectrogram-transformer#transformers.ASTConfig) configuration class: [ASTModel](/docs/transformers/v5.23.0/ko/model_doc/audio-spectrogram-transformer#transformers.ASTModel) (ASTConfig model) - [AXK1Config](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1Config) configuration class: [AXK1Model](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1Model) (AXK1Config model) - [AXK2Config](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2Config) configuration class: [AXK2Model](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2Model) (AXK2Config model) - [AfmoeConfig](/docs/transformers/v5.23.0/ko/model_doc/afmoe#transformers.AfmoeConfig) configuration class: [AfmoeModel](/docs/transformers/v5.23.0/ko/model_doc/afmoe#transformers.AfmoeModel) (AfmoeConfig model) - [Aimv2Config](/docs/transformers/v5.23.0/ko/model_doc/aimv2#transformers.Aimv2Config) configuration class: [Aimv2Model](/docs/transformers/v5.23.0/ko/model_doc/aimv2#transformers.Aimv2Model) (Aimv2Config model) - [Aimv2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/aimv2#transformers.Aimv2VisionConfig) configuration class: [Aimv2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/aimv2#transformers.Aimv2VisionModel) (Aimv2VisionConfig model) - [AlbertConfig](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertConfig) configuration class: [AlbertModel](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertModel) (AlbertConfig model) - [AlignConfig](/docs/transformers/v5.23.0/ko/model_doc/align#transformers.AlignConfig) configuration class: [AlignModel](/docs/transformers/v5.23.0/ko/model_doc/align#transformers.AlignModel) (AlignConfig model) - [AltCLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/altclip#transformers.AltCLIPConfig) configuration class: [AltCLIPModel](/docs/transformers/v5.23.0/ko/model_doc/altclip#transformers.AltCLIPModel) (AltCLIPConfig model) - [ApertusConfig](/docs/transformers/v5.23.0/ko/model_doc/apertus#transformers.ApertusConfig) configuration class: [ApertusModel](/docs/transformers/v5.23.0/ko/model_doc/apertus#transformers.ApertusModel) (ApertusConfig model) - [ArceeConfig](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeConfig) configuration class: [ArceeModel](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeModel) (ArceeConfig model) - [AriaConfig](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaConfig) configuration class: [AriaModel](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaModel) (AriaConfig model) - [AriaTextConfig](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaTextConfig) configuration class: [AriaTextModel](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaTextModel) (AriaTextConfig model) - [AudioFlamingo3Config](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3Config) configuration class: [AudioFlamingo3Model](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3Model) (AudioFlamingo3Config model) - [AudioFlamingo3EncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3EncoderConfig) configuration class: [AudioFlamingo3Encoder](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3Encoder) (AudioFlamingo3EncoderConfig model) - [AutoformerConfig](/docs/transformers/v5.23.0/ko/model_doc/autoformer#transformers.AutoformerConfig) configuration class: [AutoformerModel](/docs/transformers/v5.23.0/ko/model_doc/autoformer#transformers.AutoformerModel) (AutoformerConfig model) - [AyaVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/aya_vision#transformers.AyaVisionConfig) configuration class: [AyaVisionModel](/docs/transformers/v5.23.0/ko/model_doc/aya_vision#transformers.AyaVisionModel) (AyaVisionConfig model) - [BailingHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/bailing_hybrid#transformers.BailingHybridConfig) configuration class: [BailingHybridModel](/docs/transformers/v5.23.0/ko/model_doc/bailing_hybrid#transformers.BailingHybridModel) (BailingHybridConfig model) - [BambaConfig](/docs/transformers/v5.23.0/ko/model_doc/bamba#transformers.BambaConfig) configuration class: [BambaModel](/docs/transformers/v5.23.0/ko/model_doc/bamba#transformers.BambaModel) (BambaConfig model) - [BarkConfig](/docs/transformers/v5.23.0/ko/model_doc/bark#transformers.BarkConfig) configuration class: [BarkModel](/docs/transformers/v5.23.0/ko/model_doc/bark#transformers.BarkModel) (BarkConfig model) - [BartConfig](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartConfig) configuration class: [BartModel](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartModel) (BartConfig model) - [BeamConfig](/docs/transformers/v5.23.0/ko/model_doc/beam#transformers.BeamConfig) configuration class: [BeamModel](/docs/transformers/v5.23.0/ko/model_doc/beam#transformers.BeamModel) (BeamConfig model) - [BeitConfig](/docs/transformers/v5.23.0/ko/model_doc/beit#transformers.BeitConfig) configuration class: [BeitModel](/docs/transformers/v5.23.0/ko/model_doc/beit#transformers.BeitModel) (BeitConfig model) - [BertConfig](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertConfig) configuration class: [BertModel](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertModel) (BertConfig model) - [BertGenerationConfig](/docs/transformers/v5.23.0/ko/model_doc/bert-generation#transformers.BertGenerationConfig) configuration class: [BertGenerationEncoder](/docs/transformers/v5.23.0/ko/model_doc/bert-generation#transformers.BertGenerationEncoder) (BertGenerationConfig model) - [BigBirdConfig](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdConfig) configuration class: [BigBirdModel](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdModel) (BigBirdConfig model) - [BigBirdPegasusConfig](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusConfig) configuration class: [BigBirdPegasusModel](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusModel) (BigBirdPegasusConfig model) - [BioGptConfig](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptConfig) configuration class: [BioGptModel](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptModel) (BioGptConfig model) - [BitConfig](/docs/transformers/v5.23.0/ko/model_doc/bit#transformers.BitConfig) configuration class: [BitModel](/docs/transformers/v5.23.0/ko/model_doc/bit#transformers.BitModel) (BitConfig model) - [BitNetConfig](/docs/transformers/v5.23.0/ko/model_doc/bitnet#transformers.BitNetConfig) configuration class: [BitNetModel](/docs/transformers/v5.23.0/ko/model_doc/bitnet#transformers.BitNetModel) (BitNetConfig model) - [BlenderbotConfig](/docs/transformers/v5.23.0/ko/model_doc/blenderbot#transformers.BlenderbotConfig) configuration class: [BlenderbotModel](/docs/transformers/v5.23.0/ko/model_doc/blenderbot#transformers.BlenderbotModel) (BlenderbotConfig model) - [BlenderbotSmallConfig](/docs/transformers/v5.23.0/ko/model_doc/blenderbot-small#transformers.BlenderbotSmallConfig) configuration class: [BlenderbotSmallModel](/docs/transformers/v5.23.0/ko/model_doc/blenderbot-small#transformers.BlenderbotSmallModel) (BlenderbotSmallConfig model) - [Blip2Config](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2Config) configuration class: [Blip2Model](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2Model) (Blip2Config model) - [Blip2QFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2QFormerConfig) configuration class: [Blip2QFormerModel](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2QFormerModel) (Blip2QFormerConfig model) - [BlipConfig](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipConfig) configuration class: [BlipModel](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipModel) (BlipConfig model) - [BloomConfig](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomConfig) configuration class: [BloomModel](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomModel) (BloomConfig model) - [BltConfig](/docs/transformers/v5.23.0/ko/model_doc/blt#transformers.BltConfig) configuration class: [BltModel](/docs/transformers/v5.23.0/ko/model_doc/blt#transformers.BltModel) (BltConfig model) - [BridgeTowerConfig](/docs/transformers/v5.23.0/ko/model_doc/bridgetower#transformers.BridgeTowerConfig) configuration class: [BridgeTowerModel](/docs/transformers/v5.23.0/ko/model_doc/bridgetower#transformers.BridgeTowerModel) (BridgeTowerConfig model) - [BrosConfig](/docs/transformers/v5.23.0/ko/model_doc/bros#transformers.BrosConfig) configuration class: [BrosModel](/docs/transformers/v5.23.0/ko/model_doc/bros#transformers.BrosModel) (BrosConfig model) - [CLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPConfig) configuration class: [CLIPModel](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPModel) (CLIPConfig model) - [CLIPSegConfig](/docs/transformers/v5.23.0/ko/model_doc/clipseg#transformers.CLIPSegConfig) configuration class: [CLIPSegModel](/docs/transformers/v5.23.0/ko/model_doc/clipseg#transformers.CLIPSegModel) (CLIPSegConfig model) - [CLIPTextConfig](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTextConfig) configuration class: [CLIPTextModel](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTextModel) (CLIPTextConfig model) - [CLIPVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPVisionConfig) configuration class: [CLIPVisionModel](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPVisionModel) (CLIPVisionConfig model) - [CTRLConfig](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLConfig) configuration class: [CTRLModel](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLModel) (CTRLConfig model) - [CamembertConfig](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertConfig) configuration class: [CamembertModel](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertModel) (CamembertConfig model) - [CanaryConfig](/docs/transformers/v5.23.0/ko/model_doc/canary#transformers.CanaryConfig) configuration class: [CanaryModel](/docs/transformers/v5.23.0/ko/model_doc/canary#transformers.CanaryModel) (CanaryConfig model) - [CanineConfig](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineConfig) configuration class: [CanineModel](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineModel) (CanineConfig model) - [ChameleonConfig](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonConfig) configuration class: [ChameleonModel](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonModel) (ChameleonConfig model) - [ChineseCLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPConfig) configuration class: [ChineseCLIPModel](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPModel) (ChineseCLIPConfig model) - [ChineseCLIPVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPVisionConfig) configuration class: [ChineseCLIPVisionModel](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPVisionModel) (ChineseCLIPVisionConfig model) - [ClapConfig](/docs/transformers/v5.23.0/ko/model_doc/clap#transformers.ClapConfig) configuration class: [ClapModel](/docs/transformers/v5.23.0/ko/model_doc/clap#transformers.ClapModel) (ClapConfig model) - [ClvpConfig](/docs/transformers/v5.23.0/ko/model_doc/clvp#transformers.ClvpConfig) configuration class: [ClvpModelForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/clvp#transformers.ClvpModelForConditionalGeneration) (ClvpConfig model) - [CodeGenConfig](/docs/transformers/v5.23.0/ko/model_doc/codegen#transformers.CodeGenConfig) configuration class: [CodeGenModel](/docs/transformers/v5.23.0/ko/model_doc/codegen#transformers.CodeGenModel) (CodeGenConfig model) - [Cohere2Config](/docs/transformers/v5.23.0/ko/model_doc/cohere2#transformers.Cohere2Config) configuration class: [Cohere2Model](/docs/transformers/v5.23.0/ko/model_doc/cohere2#transformers.Cohere2Model) (Cohere2Config model) - [Cohere2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere2_moe#transformers.Cohere2MoeConfig) configuration class: [Cohere2MoeModel](/docs/transformers/v5.23.0/ko/model_doc/cohere2_moe#transformers.Cohere2MoeModel) (Cohere2MoeConfig model) - [Cohere2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere2_vision#transformers.Cohere2VisionConfig) configuration class: [Cohere2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/cohere2_vision#transformers.Cohere2VisionModel) (Cohere2VisionConfig model) - [CohereAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_asr#transformers.CohereAsrConfig) configuration class: [CohereAsrModel](/docs/transformers/v5.23.0/ko/model_doc/cohere_asr#transformers.CohereAsrModel) (CohereAsrConfig model) - [CohereCompassConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassConfig) configuration class: [CohereCompassModel](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassModel) (CohereCompassConfig model) - [CohereCompassTextConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassTextConfig) configuration class: [CohereCompassTextModel](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassTextModel) (CohereCompassTextConfig model) - [CohereCompassVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassVisionConfig) configuration class: [CohereCompassVisionModel](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassVisionModel) (CohereCompassVisionConfig model) - [CohereConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere#transformers.CohereConfig) configuration class: [CohereModel](/docs/transformers/v5.23.0/ko/model_doc/cohere#transformers.CohereModel) (CohereConfig model) - [ConditionalDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/conditional_detr#transformers.ConditionalDetrConfig) configuration class: [ConditionalDetrModel](/docs/transformers/v5.23.0/ko/model_doc/conditional_detr#transformers.ConditionalDetrModel) (ConditionalDetrConfig model) - [ConvBertConfig](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertConfig) configuration class: [ConvBertModel](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertModel) (ConvBertConfig model) - [ConvNextConfig](/docs/transformers/v5.23.0/ko/model_doc/convnext#transformers.ConvNextConfig) configuration class: [ConvNextModel](/docs/transformers/v5.23.0/ko/model_doc/convnext#transformers.ConvNextModel) (ConvNextConfig model) - [ConvNextV2Config](/docs/transformers/v5.23.0/ko/model_doc/convnextv2#transformers.ConvNextV2Config) configuration class: [ConvNextV2Model](/docs/transformers/v5.23.0/ko/model_doc/convnextv2#transformers.ConvNextV2Model) (ConvNextV2Config model) - [Cosmos3EdgeConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeConfig) configuration class: [Cosmos3EdgeModel](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeModel) (Cosmos3EdgeConfig model) - [Cosmos3EdgeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeTextConfig) configuration class: [Cosmos3EdgeTextModel](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeTextModel) (Cosmos3EdgeTextConfig model) - [Cosmos3EdgeVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeVisionConfig) configuration class: [Cosmos3EdgeVisionModel](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeVisionModel) (Cosmos3EdgeVisionConfig model) - [Cosmos3OmniConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_omni#transformers.Cosmos3OmniConfig) configuration class: [Cosmos3OmniModel](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_omni#transformers.Cosmos3OmniModel) (Cosmos3OmniConfig model) - [CpmAntConfig](/docs/transformers/v5.23.0/ko/model_doc/cpmant#transformers.CpmAntConfig) configuration class: [CpmAntModel](/docs/transformers/v5.23.0/ko/model_doc/cpmant#transformers.CpmAntModel) (CpmAntConfig model) - [CsmConfig](/docs/transformers/v5.23.0/ko/model_doc/csm#transformers.CsmConfig) configuration class: [CsmForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/csm#transformers.CsmForConditionalGeneration) (CsmConfig model) - [CvtConfig](/docs/transformers/v5.23.0/ko/model_doc/cvt#transformers.CvtConfig) configuration class: [CvtModel](/docs/transformers/v5.23.0/ko/model_doc/cvt#transformers.CvtModel) (CvtConfig model) - [CwmConfig](/docs/transformers/v5.23.0/ko/model_doc/cwm#transformers.CwmConfig) configuration class: [CwmModel](/docs/transformers/v5.23.0/ko/model_doc/cwm#transformers.CwmModel) (CwmConfig model) - [DFineConfig](/docs/transformers/v5.23.0/ko/model_doc/d_fine#transformers.DFineConfig) configuration class: [DFineModel](/docs/transformers/v5.23.0/ko/model_doc/d_fine#transformers.DFineModel) (DFineConfig model) - [DINOv3ConvNextConfig](/docs/transformers/v5.23.0/ko/model_doc/dinov3#transformers.DINOv3ConvNextConfig) configuration class: [DINOv3ConvNextModel](/docs/transformers/v5.23.0/ko/model_doc/dinov3#transformers.DINOv3ConvNextModel) (DINOv3ConvNextConfig model) - [DINOv3ViTConfig](/docs/transformers/v5.23.0/ko/model_doc/dinov3#transformers.DINOv3ViTConfig) configuration class: [DINOv3ViTModel](/docs/transformers/v5.23.0/ko/model_doc/dinov3#transformers.DINOv3ViTModel) (DINOv3ViTConfig model) - [DPRConfig](/docs/transformers/v5.23.0/ko/model_doc/dpr#transformers.DPRConfig) configuration class: [DPRQuestionEncoder](/docs/transformers/v5.23.0/ko/model_doc/dpr#transformers.DPRQuestionEncoder) (DPRConfig model) - [DPTConfig](/docs/transformers/v5.23.0/ko/model_doc/dpt#transformers.DPTConfig) configuration class: [DPTModel](/docs/transformers/v5.23.0/ko/model_doc/dpt#transformers.DPTModel) (DPTConfig model) - [DabDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/dab-detr#transformers.DabDetrConfig) configuration class: [DabDetrModel](/docs/transformers/v5.23.0/ko/model_doc/dab-detr#transformers.DabDetrModel) (DabDetrConfig model) - [DacConfig](/docs/transformers/v5.23.0/ko/model_doc/dac#transformers.DacConfig) configuration class: [DacModel](/docs/transformers/v5.23.0/ko/model_doc/dac#transformers.DacModel) (DacConfig model) - [Data2VecAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioConfig) configuration class: [Data2VecAudioModel](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioModel) (Data2VecAudioConfig model) - [Data2VecTextConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextConfig) configuration class: [Data2VecTextModel](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextModel) (Data2VecTextConfig model) - [Data2VecVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecVisionConfig) configuration class: [Data2VecVisionModel](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecVisionModel) (Data2VecVisionConfig model) - [DbrxConfig](/docs/transformers/v5.23.0/ko/model_doc/dbrx#transformers.DbrxConfig) configuration class: [DbrxModel](/docs/transformers/v5.23.0/ko/model_doc/dbrx#transformers.DbrxModel) (DbrxConfig model) - [DebertaConfig](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaConfig) configuration class: [DebertaModel](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaModel) (DebertaConfig model) - [DebertaV2Config](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2Config) configuration class: [DebertaV2Model](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2Model) (DebertaV2Config model) - [DecisionTransformerConfig](/docs/transformers/v5.23.0/ko/model_doc/decision_transformer#transformers.DecisionTransformerConfig) configuration class: [DecisionTransformerModel](/docs/transformers/v5.23.0/ko/model_doc/decision_transformer#transformers.DecisionTransformerModel) (DecisionTransformerConfig model) - [DeepseekOcr2Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2Config) configuration class: [DeepseekOcr2Model](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2Model) (DeepseekOcr2Config model) - [DeepseekV2Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v2#transformers.DeepseekV2Config) configuration class: [DeepseekV2Model](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v2#transformers.DeepseekV2Model) (DeepseekV2Config model) - [DeepseekV32Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v32#transformers.DeepseekV32Config) configuration class: [DeepseekV32Model](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v32#transformers.DeepseekV32Model) (DeepseekV32Config model) - [DeepseekV3Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3Config) configuration class: [DeepseekV3Model](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3Model) (DeepseekV3Config model) - [DeepseekV4Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v4#transformers.DeepseekV4Config) configuration class: [DeepseekV4Model](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v4#transformers.DeepseekV4Model) (DeepseekV4Config model) - [DeepseekVLConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl#transformers.DeepseekVLConfig) configuration class: [DeepseekVLModel](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl#transformers.DeepseekVLModel) (DeepseekVLConfig model) - [DeepseekVLHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl_hybrid#transformers.DeepseekVLHybridConfig) configuration class: [DeepseekVLHybridModel](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl_hybrid#transformers.DeepseekVLHybridModel) (DeepseekVLHybridConfig model) - [DeformableDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/deformable_detr#transformers.DeformableDetrConfig) configuration class: [DeformableDetrModel](/docs/transformers/v5.23.0/ko/model_doc/deformable_detr#transformers.DeformableDetrModel) (DeformableDetrConfig model) - [DeiTConfig](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTConfig) configuration class: [DeiTModel](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTModel) (DeiTConfig model) - [Deimv2Config](/docs/transformers/v5.23.0/ko/model_doc/deimv2#transformers.Deimv2Config) configuration class: [Deimv2Model](/docs/transformers/v5.23.0/ko/model_doc/deimv2#transformers.Deimv2Model) (Deimv2Config model) - [DepthProConfig](/docs/transformers/v5.23.0/ko/model_doc/depth_pro#transformers.DepthProConfig) configuration class: [DepthProModel](/docs/transformers/v5.23.0/ko/model_doc/depth_pro#transformers.DepthProModel) (DepthProConfig model) - [DetrConfig](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrConfig) configuration class: [DetrModel](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrModel) (DetrConfig model) - [DiaConfig](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaConfig) configuration class: [DiaModel](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaModel) (DiaConfig model) - [DiffLlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaConfig) configuration class: [DiffLlamaModel](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaModel) (DiffLlamaConfig model) - [DiffusionGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/diffusion_gemma#transformers.DiffusionGemmaConfig) configuration class: [DiffusionGemmaModel](/docs/transformers/v5.23.0/ko/model_doc/diffusion_gemma#transformers.DiffusionGemmaModel) (DiffusionGemmaConfig model) - [DinatConfig](/docs/transformers/v5.23.0/ko/model_doc/dinat#transformers.DinatConfig) configuration class: [DinatModel](/docs/transformers/v5.23.0/ko/model_doc/dinat#transformers.DinatModel) (DinatConfig model) - [Dinov2Config](/docs/transformers/v5.23.0/ko/model_doc/dinov2#transformers.Dinov2Config) configuration class: [Dinov2Model](/docs/transformers/v5.23.0/ko/model_doc/dinov2#transformers.Dinov2Model) (Dinov2Config model) - [Dinov2WithRegistersConfig](/docs/transformers/v5.23.0/ko/model_doc/dinov2_with_registers#transformers.Dinov2WithRegistersConfig) configuration class: [Dinov2WithRegistersModel](/docs/transformers/v5.23.0/ko/model_doc/dinov2_with_registers#transformers.Dinov2WithRegistersModel) (Dinov2WithRegistersConfig model) - [DistilBertConfig](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertConfig) configuration class: [DistilBertModel](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertModel) (DistilBertConfig model) - [DogeConfig](/docs/transformers/v5.23.0/ko/model_doc/doge#transformers.DogeConfig) configuration class: [DogeModel](/docs/transformers/v5.23.0/ko/model_doc/doge#transformers.DogeModel) (DogeConfig model) - [DonutSwinConfig](/docs/transformers/v5.23.0/ko/model_doc/donut#transformers.DonutSwinConfig) configuration class: [DonutSwinModel](/docs/transformers/v5.23.0/ko/model_doc/donut#transformers.DonutSwinModel) (DonutSwinConfig model) - [Dots1Config](/docs/transformers/v5.23.0/ko/model_doc/dots1#transformers.Dots1Config) configuration class: [Dots1Model](/docs/transformers/v5.23.0/ko/model_doc/dots1#transformers.Dots1Model) (Dots1Config model) - [Dots3NoteConfig](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteConfig) configuration class: [Dots3NoteModel](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteModel) (Dots3NoteConfig model) - [Dots3NoteTextConfig](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteTextConfig) configuration class: [Dots3NoteTextModel](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteTextModel) (Dots3NoteTextConfig model) - [EdgeTamConfig](/docs/transformers/v5.23.0/ko/model_doc/edgetam#transformers.EdgeTamConfig) configuration class: [EdgeTamModel](/docs/transformers/v5.23.0/ko/model_doc/edgetam#transformers.EdgeTamModel) (EdgeTamConfig model) - [EdgeTamVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/edgetam_video#transformers.EdgeTamVideoConfig) configuration class: [EdgeTamVideoModel](/docs/transformers/v5.23.0/ko/model_doc/edgetam_video#transformers.EdgeTamVideoModel) (EdgeTamVideoConfig model) - [EdgeTamVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/edgetam#transformers.EdgeTamVisionConfig) configuration class: [EdgeTamVisionModel](/docs/transformers/v5.23.0/ko/model_doc/edgetam#transformers.EdgeTamVisionModel) (EdgeTamVisionConfig model) - [EfficientLoFTRConfig](/docs/transformers/v5.23.0/ko/model_doc/efficientloftr#transformers.EfficientLoFTRConfig) configuration class: [EfficientLoFTRModel](/docs/transformers/v5.23.0/ko/model_doc/efficientloftr#transformers.EfficientLoFTRModel) (EfficientLoFTRConfig model) - [EfficientNetConfig](/docs/transformers/v5.23.0/ko/model_doc/efficientnet#transformers.EfficientNetConfig) configuration class: [EfficientNetModel](/docs/transformers/v5.23.0/ko/model_doc/efficientnet#transformers.EfficientNetModel) (EfficientNetConfig model) - [ElectraConfig](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraConfig) configuration class: [ElectraModel](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraModel) (ElectraConfig model) - [EmbeddingGemma2Config](/docs/transformers/v5.23.0/ko/model_doc/embedding_gemma2#transformers.EmbeddingGemma2Config) configuration class: [EmbeddingGemma2Model](/docs/transformers/v5.23.0/ko/model_doc/embedding_gemma2#transformers.EmbeddingGemma2Model) (EmbeddingGemma2Config model) - [EmbeddingGemma2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/embedding_gemma2#transformers.EmbeddingGemma2TextConfig) configuration class: [EmbeddingGemma2TextModel](/docs/transformers/v5.23.0/ko/model_doc/embedding_gemma2#transformers.EmbeddingGemma2TextModel) (EmbeddingGemma2TextConfig model) - [Emu3Config](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3Config) configuration class: [Emu3Model](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3Model) (Emu3Config model) - [EncodecConfig](/docs/transformers/v5.23.0/ko/model_doc/encodec#transformers.EncodecConfig) configuration class: [EncodecModel](/docs/transformers/v5.23.0/ko/model_doc/encodec#transformers.EncodecModel) (EncodecConfig model) - [Ernie4_5Config](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5#transformers.Ernie4_5Config) configuration class: [Ernie4_5Model](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5#transformers.Ernie4_5Model) (Ernie4_5Config model) - [Ernie4_5_MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_moe#transformers.Ernie4_5_MoeConfig) configuration class: [Ernie4_5_MoeModel](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_moe#transformers.Ernie4_5_MoeModel) (Ernie4_5_MoeConfig model) - [Ernie4_5_VLMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeConfig) configuration class: [Ernie4_5_VLMoeModel](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeModel) (Ernie4_5_VLMoeConfig model) - [ErnieConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieConfig) configuration class: [ErnieModel](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieModel) (ErnieConfig model) - [EsmConfig](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmConfig) configuration class: [EsmModel](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmModel) (EsmConfig model) - [EsmcConfig](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcConfig) configuration class: [EsmcModel](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcModel) (EsmcConfig model) - [EuroBertConfig](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertConfig) configuration class: [EuroBertModel](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertModel) (EuroBertConfig model) - [EvollaConfig](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaConfig) configuration class: [EvollaModel](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaModel) (EvollaConfig model) - [Exaone4Config](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4Config) configuration class: [Exaone4Model](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4Model) (Exaone4Config model) - [Exaone4_5_Config](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_Config) configuration class: [Exaone4_5_Model](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_Model) (Exaone4_5_Config model) - [Exaone4_5_VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_VisionConfig) configuration class: [Exaone4_5_VisionModel](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_VisionModel) (Exaone4_5_VisionConfig model) - [ExaoneMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/exaone_moe#transformers.ExaoneMoeConfig) configuration class: [ExaoneMoeModel](/docs/transformers/v5.23.0/ko/model_doc/exaone_moe#transformers.ExaoneMoeModel) (ExaoneMoeConfig model) - [FNetConfig](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetConfig) configuration class: [FNetModel](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetModel) (FNetConfig model) - [FSMTConfig](/docs/transformers/v5.23.0/ko/model_doc/fsmt#transformers.FSMTConfig) configuration class: [FSMTModel](/docs/transformers/v5.23.0/ko/model_doc/fsmt#transformers.FSMTModel) (FSMTConfig model) - [FalconConfig](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconConfig) configuration class: [FalconModel](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconModel) (FalconConfig model) - [FalconH1Config](/docs/transformers/v5.23.0/ko/model_doc/falcon_h1#transformers.FalconH1Config) configuration class: [FalconH1Model](/docs/transformers/v5.23.0/ko/model_doc/falcon_h1#transformers.FalconH1Model) (FalconH1Config model) - [FalconMambaConfig](/docs/transformers/v5.23.0/ko/model_doc/falcon_mamba#transformers.FalconMambaConfig) configuration class: [FalconMambaModel](/docs/transformers/v5.23.0/ko/model_doc/falcon_mamba#transformers.FalconMambaModel) (FalconMambaConfig model) - [FastSpeech2ConformerConfig](/docs/transformers/v5.23.0/ko/model_doc/fastspeech2_conformer#transformers.FastSpeech2ConformerConfig) configuration class: [FastSpeech2ConformerModel](/docs/transformers/v5.23.0/ko/model_doc/fastspeech2_conformer#transformers.FastSpeech2ConformerModel) (FastSpeech2ConformerConfig model) - [FastSpeech2ConformerWithHifiGanConfig](/docs/transformers/v5.23.0/ko/model_doc/fastspeech2_conformer#transformers.FastSpeech2ConformerWithHifiGanConfig) configuration class: [FastSpeech2ConformerWithHifiGan](/docs/transformers/v5.23.0/ko/model_doc/fastspeech2_conformer#transformers.FastSpeech2ConformerWithHifiGan) (FastSpeech2ConformerWithHifiGanConfig model) - [FastVlmConfig](/docs/transformers/v5.23.0/ko/model_doc/fast_vlm#transformers.FastVlmConfig) configuration class: [FastVlmModel](/docs/transformers/v5.23.0/ko/model_doc/fast_vlm#transformers.FastVlmModel) (FastVlmConfig model) - [FlaubertConfig](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertConfig) configuration class: [FlaubertModel](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertModel) (FlaubertConfig model) - [FlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaConfig) configuration class: [FlavaModel](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaModel) (FlavaConfig model) - [FlexOlmoConfig](/docs/transformers/v5.23.0/ko/model_doc/flex_olmo#transformers.FlexOlmoConfig) configuration class: [FlexOlmoModel](/docs/transformers/v5.23.0/ko/model_doc/flex_olmo#transformers.FlexOlmoModel) (FlexOlmoConfig model) - [FlintConfig](/docs/transformers/v5.23.0/ko/model_doc/flint#transformers.FlintConfig) configuration class: [FlintModel](/docs/transformers/v5.23.0/ko/model_doc/flint#transformers.FlintModel) (FlintConfig model) - [Florence2Config](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2Config) configuration class: [Florence2Model](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2Model) (Florence2Config model) - [FocalNetConfig](/docs/transformers/v5.23.0/ko/model_doc/focalnet#transformers.FocalNetConfig) configuration class: [FocalNetModel](/docs/transformers/v5.23.0/ko/model_doc/focalnet#transformers.FocalNetModel) (FocalNetConfig model) - [FunAsrNanoConfig](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoConfig) configuration class: [FunAsrNanoModel](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoModel) (FunAsrNanoConfig model) - [FunAsrNanoEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoEncoderConfig) configuration class: [FunAsrNanoEncoder](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoEncoder) (FunAsrNanoEncoderConfig model) - [FunnelConfig](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelConfig) configuration class: [FunnelModel](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelModel) or [FunnelBaseModel](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelBaseModel) (FunnelConfig model) - [FuyuConfig](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuConfig) configuration class: [FuyuModel](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuModel) (FuyuConfig model) - [GLPNConfig](/docs/transformers/v5.23.0/ko/model_doc/glpn#transformers.GLPNConfig) configuration class: [GLPNModel](/docs/transformers/v5.23.0/ko/model_doc/glpn#transformers.GLPNModel) (GLPNConfig model) - [GPT2Config](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Config) configuration class: [GPT2Model](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Model) (GPT2Config model) - [GPTBigCodeConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeConfig) configuration class: [GPTBigCodeModel](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeModel) (GPTBigCodeConfig model) - [GPTJConfig](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJConfig) configuration class: [GPTJModel](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJModel) (GPTJConfig model) - [GPTNeoConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoConfig) configuration class: [GPTNeoModel](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoModel) (GPTNeoConfig model) - [GPTNeoXConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXConfig) configuration class: [GPTNeoXModel](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXModel) (GPTNeoXConfig model) - [GPTNeoXJapaneseConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox_japanese#transformers.GPTNeoXJapaneseConfig) configuration class: [GPTNeoXJapaneseModel](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox_japanese#transformers.GPTNeoXJapaneseModel) (GPTNeoXJapaneseConfig model) - [Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2Config) configuration class: [Gemma2Model](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2Model) (Gemma2Config model) - [Gemma3Config](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Config) configuration class: [Gemma3Model](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Model) (Gemma3Config model) - [Gemma3TextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3TextConfig) configuration class: [Gemma3TextModel](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3TextModel) (Gemma3TextConfig model) - [Gemma3nAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nAudioConfig) configuration class: `Gemma3nAudioEncoder` (Gemma3nAudioConfig model) - [Gemma3nConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nConfig) configuration class: [Gemma3nModel](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nModel) (Gemma3nConfig model) - [Gemma3nTextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nTextConfig) configuration class: [Gemma3nTextModel](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nTextModel) (Gemma3nTextConfig model) - [Gemma3nVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nVisionConfig) configuration class: [TimmWrapperModel](/docs/transformers/v5.23.0/ko/model_doc/timm_wrapper#transformers.TimmWrapperModel) (Gemma3nVisionConfig model) - [Gemma4AudioConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4AudioConfig) configuration class: [Gemma4AudioModel](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4AudioModel) (Gemma4AudioConfig model) - [Gemma4Config](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4Config) configuration class: [Gemma4Model](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4Model) (Gemma4Config model) - [Gemma4TextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4TextConfig) configuration class: [Gemma4TextModel](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4TextModel) (Gemma4TextConfig model) - [Gemma4UnifiedConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedConfig) configuration class: [Gemma4UnifiedModel](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedModel) (Gemma4UnifiedConfig model) - [Gemma4UnifiedTextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedTextConfig) configuration class: [Gemma4UnifiedTextModel](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedTextModel) (Gemma4UnifiedTextConfig model) - [Gemma4VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4VisionConfig) configuration class: [Gemma4VisionModel](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4VisionModel) (Gemma4VisionConfig model) - [GemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaConfig) configuration class: [GemmaModel](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaModel) (GemmaConfig model) - [GitConfig](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitConfig) configuration class: [GitModel](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitModel) (GitConfig model) - [Glm46VConfig](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VConfig) configuration class: [Glm46VModel](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VModel) (Glm46VConfig model) - [Glm4Config](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4Config) configuration class: [Glm4Model](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4Model) (Glm4Config model) - [Glm4MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe#transformers.Glm4MoeConfig) configuration class: [Glm4MoeModel](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe#transformers.Glm4MoeModel) (Glm4MoeConfig model) - [Glm4MoeLiteConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe_lite#transformers.Glm4MoeLiteConfig) configuration class: [Glm4MoeLiteModel](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe_lite#transformers.Glm4MoeLiteModel) (Glm4MoeLiteConfig model) - [Glm4vConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vConfig) configuration class: [Glm4vModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vModel) (Glm4vConfig model) - [Glm4vMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeConfig) configuration class: [Glm4vMoeModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeModel) (Glm4vMoeConfig model) - [Glm4vMoeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeTextConfig) configuration class: [Glm4vMoeTextModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeTextModel) (Glm4vMoeTextConfig model) - [Glm4vMoeVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeVisionConfig) configuration class: [Glm4vMoeVisionModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeVisionModel) (Glm4vMoeVisionConfig model) - [Glm4vTextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vTextConfig) configuration class: [Glm4vTextModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vTextModel) (Glm4vTextConfig model) - [Glm4vVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vVisionConfig) configuration class: [Glm4vVisionModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vVisionModel) (Glm4vVisionConfig model) - [Glm5NextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextConfig) configuration class: [Glm5NextModel](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextModel) (Glm5NextConfig model) - [Glm5NextTextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextTextConfig) configuration class: [Glm5NextTextModel](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextTextModel) (Glm5NextTextConfig model) - [Glm5NextVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextVisionConfig) configuration class: [Glm5NextVisionModel](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextVisionModel) (Glm5NextVisionConfig model) - [GlmAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrConfig) configuration class: [GlmAsrModel](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrModel) (GlmAsrConfig model) - [GlmAsrEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrEncoderConfig) configuration class: [GlmAsrEncoder](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrEncoder) (GlmAsrEncoderConfig model) - [GlmConfig](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmConfig) configuration class: [GlmModel](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmModel) (GlmConfig model) - [GlmImageConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageConfig) configuration class: [GlmImageModel](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageModel) (GlmImageConfig model) - [GlmImageTextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageTextConfig) configuration class: [GlmImageTextModel](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageTextModel) (GlmImageTextConfig model) - [GlmImageVQVAEConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageVQVAEConfig) configuration class: [GlmImageVQVAE](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageVQVAE) (GlmImageVQVAEConfig model) - [GlmImageVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageVisionConfig) configuration class: [GlmImageVisionModel](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageVisionModel) (GlmImageVisionConfig model) - [GlmMoeDsaConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_moe_dsa#transformers.GlmMoeDsaConfig) configuration class: [GlmMoeDsaModel](/docs/transformers/v5.23.0/ko/model_doc/glm_moe_dsa#transformers.GlmMoeDsaModel) (GlmMoeDsaConfig model) - [GlmOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrConfig) configuration class: [GlmOcrModel](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrModel) (GlmOcrConfig model) - [GlmOcrTextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrTextConfig) configuration class: [GlmOcrTextModel](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrTextModel) (GlmOcrTextConfig model) - [GlmOcrVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrVisionConfig) configuration class: [GlmOcrVisionModel](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrVisionModel) (GlmOcrVisionConfig model) - [GlmgaConfig](/docs/transformers/v5.23.0/ko/model_doc/glmga#transformers.GlmgaConfig) configuration class: [Glm46VModel](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VModel) (GlmgaConfig model) - [GotOcr2Config](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2Config) configuration class: [GotOcr2Model](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2Model) (GotOcr2Config model) - [GptOssConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssConfig) configuration class: [GptOssModel](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssModel) (GptOssConfig model) - [GradiumPhononTTSConfig](/docs/transformers/v5.23.0/ko/model_doc/gradium_phonon_tts#transformers.GradiumPhononTTSConfig) configuration class: [GradiumPhononTTSModel](/docs/transformers/v5.23.0/ko/model_doc/gradium_phonon_tts#transformers.GradiumPhononTTSModel) (GradiumPhononTTSConfig model) - [Granite4VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionConfig) configuration class: [Granite4VisionModel](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionModel) (Granite4VisionConfig model) - [GraniteConfig](/docs/transformers/v5.23.0/ko/model_doc/granite#transformers.GraniteConfig) configuration class: [GraniteModel](/docs/transformers/v5.23.0/ko/model_doc/granite#transformers.GraniteModel) (GraniteConfig model) - [GraniteDoclingConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingConfig) configuration class: [GraniteDoclingModel](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingModel) (GraniteDoclingConfig model) - [GraniteDoclingTextConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingTextConfig) configuration class: [GraniteDoclingTextModel](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingTextModel) (GraniteDoclingTextConfig model) - [GraniteDoclingVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingVisionConfig) configuration class: [GraniteDoclingVisionModel](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingVisionModel) (GraniteDoclingVisionConfig model) - [GraniteMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoe#transformers.GraniteMoeConfig) configuration class: [GraniteMoeModel](/docs/transformers/v5.23.0/ko/model_doc/granitemoe#transformers.GraniteMoeModel) (GraniteMoeConfig model) - [GraniteMoeHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoehybrid#transformers.GraniteMoeHybridConfig) configuration class: [GraniteMoeHybridModel](/docs/transformers/v5.23.0/ko/model_doc/granitemoehybrid#transformers.GraniteMoeHybridModel) (GraniteMoeHybridConfig model) - [GraniteMoeSWAConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoe_swa#transformers.GraniteMoeSWAConfig) configuration class: [GraniteMoeSWAModel](/docs/transformers/v5.23.0/ko/model_doc/granitemoe_swa#transformers.GraniteMoeSWAModel) (GraniteMoeSWAConfig model) - [GraniteMoeSharedConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoeshared#transformers.GraniteMoeSharedConfig) configuration class: [GraniteMoeSharedModel](/docs/transformers/v5.23.0/ko/model_doc/granitemoeshared#transformers.GraniteMoeSharedModel) (GraniteMoeSharedConfig model) - [GraniteSWAConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_swa#transformers.GraniteSWAConfig) configuration class: [GraniteSWAModel](/docs/transformers/v5.23.0/ko/model_doc/granite_swa#transformers.GraniteSWAModel) (GraniteSWAConfig model) - [GraniteSpeech5CTCConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5CTCConfig) configuration class: [GraniteSpeech5ForCTC](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5ForCTC) (GraniteSpeech5CTCConfig model) - [GraniteSpeech5EncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5EncoderConfig) configuration class: [GraniteSpeech5Encoder](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5Encoder) (GraniteSpeech5EncoderConfig model) - [GraniteSpeechConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechConfig) configuration class: [GraniteSpeechModel](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechModel) (GraniteSpeechConfig model) - [GraniteSpeechPlusConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusConfig) configuration class: [GraniteSpeechPlusModel](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusModel) (GraniteSpeechPlusConfig model) - [GroundingDinoConfig](/docs/transformers/v5.23.0/ko/model_doc/grounding-dino#transformers.GroundingDinoConfig) configuration class: [GroundingDinoModel](/docs/transformers/v5.23.0/ko/model_doc/grounding-dino#transformers.GroundingDinoModel) (GroundingDinoConfig model) - [GroupViTConfig](/docs/transformers/v5.23.0/ko/model_doc/groupvit#transformers.GroupViTConfig) configuration class: [GroupViTModel](/docs/transformers/v5.23.0/ko/model_doc/groupvit#transformers.GroupViTModel) (GroupViTConfig model) - [GteConfig](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteConfig) configuration class: [GteModel](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteModel) (GteConfig model) - [HGNetV2Config](/docs/transformers/v5.23.0/ko/model_doc/hgnet_v2#transformers.HGNetV2Config) configuration class: [HGNetV2Backbone](/docs/transformers/v5.23.0/ko/model_doc/hgnet_v2#transformers.HGNetV2Backbone) (HGNetV2Config model) - [HYV3Config](/docs/transformers/v5.23.0/ko/model_doc/hy_v3#transformers.HYV3Config) configuration class: [HYV3Model](/docs/transformers/v5.23.0/ko/model_doc/hy_v3#transformers.HYV3Model) (HYV3Config model) - [HYV4Config](/docs/transformers/v5.23.0/ko/model_doc/hy_v4#transformers.HYV4Config) configuration class: [HYV4Model](/docs/transformers/v5.23.0/ko/model_doc/hy_v4#transformers.HYV4Model) (HYV4Config model) - [HeliumConfig](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumConfig) configuration class: [HeliumModel](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumModel) (HeliumConfig model) - [HieraConfig](/docs/transformers/v5.23.0/ko/model_doc/hiera#transformers.HieraConfig) configuration class: [HieraModel](/docs/transformers/v5.23.0/ko/model_doc/hiera#transformers.HieraModel) (HieraConfig model) - [HiggsAudioV2Config](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2#transformers.HiggsAudioV2Config) configuration class: [HiggsAudioV2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2#transformers.HiggsAudioV2ForConditionalGeneration) (HiggsAudioV2Config model) - [HiggsAudioV2TokenizerConfig](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2_tokenizer#transformers.HiggsAudioV2TokenizerConfig) configuration class: [HiggsAudioV2TokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2_tokenizer#transformers.HiggsAudioV2TokenizerModel) (HiggsAudioV2TokenizerConfig model) - [HrmTextConfig](/docs/transformers/v5.23.0/ko/model_doc/hrm_text#transformers.HrmTextConfig) configuration class: [HrmTextModel](/docs/transformers/v5.23.0/ko/model_doc/hrm_text#transformers.HrmTextModel) (HrmTextConfig model) - [HubertConfig](/docs/transformers/v5.23.0/ko/model_doc/hubert#transformers.HubertConfig) configuration class: [HubertModel](/docs/transformers/v5.23.0/ko/model_doc/hubert#transformers.HubertModel) (HubertConfig model) - [HunYuanDenseV1Config](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_dense#transformers.HunYuanDenseV1Config) configuration class: [HunYuanDenseV1Model](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_dense#transformers.HunYuanDenseV1Model) (HunYuanDenseV1Config model) - [HunYuanMoEV1Config](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_moe#transformers.HunYuanMoEV1Config) configuration class: [HunYuanMoEV1Model](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_moe#transformers.HunYuanMoEV1Model) (HunYuanMoEV1Config model) - [HunYuanVLConfig](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLConfig) configuration class: [HunYuanVLModel](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLModel) (HunYuanVLConfig model) - [HunYuanVLTextConfig](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLTextConfig) configuration class: [HunYuanVLTextModel](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLTextModel) (HunYuanVLTextConfig model) - [HunYuanVLVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLVisionConfig) configuration class: [HunYuanVLVisionTransformer](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLVisionTransformer) (HunYuanVLVisionConfig model) - [HyperCLOVAXConfig](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax#transformers.HyperCLOVAXConfig) configuration class: [HyperCLOVAXModel](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax#transformers.HyperCLOVAXModel) (HyperCLOVAXConfig model) - [HyperCLOVAXVisionV2Config](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax_vision_v2#transformers.HyperCLOVAXVisionV2Config) configuration class: [HyperCLOVAXVisionV2Model](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax_vision_v2#transformers.HyperCLOVAXVisionV2Model) (HyperCLOVAXVisionV2Config model) - [IBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertConfig) configuration class: [IBertModel](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertModel) (IBertConfig model) - [IJepaConfig](/docs/transformers/v5.23.0/ko/model_doc/ijepa#transformers.IJepaConfig) configuration class: [IJepaModel](/docs/transformers/v5.23.0/ko/model_doc/ijepa#transformers.IJepaModel) (IJepaConfig model) - [IQuestCoderConfig](/docs/transformers/v5.23.0/ko/model_doc/iquestcoder#transformers.IQuestCoderConfig) configuration class: [LlamaModel](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaModel) (IQuestCoderConfig model) - [Idefics2Config](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2Config) configuration class: [Idefics2Model](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2Model) (Idefics2Config model) - [Idefics3Config](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3Config) configuration class: [Idefics3Model](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3Model) (Idefics3Config model) - [Idefics3VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3VisionConfig) configuration class: [Idefics3VisionTransformer](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3VisionTransformer) (Idefics3VisionConfig model) - [IdeficsConfig](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsConfig) configuration class: [IdeficsModel](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsModel) (IdeficsConfig model) - [ImageGPTConfig](/docs/transformers/v5.23.0/ko/model_doc/imagegpt#transformers.ImageGPTConfig) configuration class: [ImageGPTModel](/docs/transformers/v5.23.0/ko/model_doc/imagegpt#transformers.ImageGPTModel) (ImageGPTConfig model) - [InformerConfig](/docs/transformers/v5.23.0/ko/model_doc/informer#transformers.InformerConfig) configuration class: [InformerModel](/docs/transformers/v5.23.0/ko/model_doc/informer#transformers.InformerModel) (InformerConfig model) - [Inkling2Config](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2Config) configuration class: [Inkling2Model](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2Model) (Inkling2Config model) - [Inkling2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2TextConfig) configuration class: [Inkling2TextModel](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2TextModel) (Inkling2TextConfig model) - [Inkling2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2VisionConfig) configuration class: [Inkling2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2VisionModel) (Inkling2VisionConfig model) - [InklingConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingConfig) configuration class: [InklingModel](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingModel) (InklingConfig model) - [InklingTextConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingTextConfig) configuration class: [InklingTextModel](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingTextModel) (InklingTextConfig model) - [InstructBlipConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipConfig) configuration class: [InstructBlipModel](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipModel) (InstructBlipConfig model) - [InstructBlipVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoConfig) configuration class: [InstructBlipVideoModel](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoModel) (InstructBlipVideoConfig model) - [InternLM3Config](/docs/transformers/v5.23.0/ko/model_doc/internlm3#transformers.InternLM3Config) configuration class: [LlamaModel](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaModel) (InternLM3Config model) - [InternVLConfig](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLConfig) configuration class: [InternVLModel](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLModel) (InternVLConfig model) - [InternVLVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLVisionConfig) configuration class: [InternVLVisionModel](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLVisionModel) (InternVLVisionConfig model) - [Jais2Config](/docs/transformers/v5.23.0/ko/model_doc/jais2#transformers.Jais2Config) configuration class: [Jais2Model](/docs/transformers/v5.23.0/ko/model_doc/jais2#transformers.Jais2Model) (Jais2Config model) - [JambaConfig](/docs/transformers/v5.23.0/ko/model_doc/jamba#transformers.JambaConfig) configuration class: [JambaModel](/docs/transformers/v5.23.0/ko/model_doc/jamba#transformers.JambaModel) (JambaConfig model) - [JanusConfig](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusConfig) configuration class: [JanusModel](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusModel) (JanusConfig model) - [JetMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/jetmoe#transformers.JetMoeConfig) configuration class: [JetMoeModel](/docs/transformers/v5.23.0/ko/model_doc/jetmoe#transformers.JetMoeModel) (JetMoeConfig model) - [JinaEmbeddingsV3Config](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3Config) configuration class: [JinaEmbeddingsV3Model](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3Model) (JinaEmbeddingsV3Config model) - [KimiLinearConfig](/docs/transformers/v5.23.0/ko/model_doc/kimi_linear#transformers.KimiLinearConfig) configuration class: [KimiLinearModel](/docs/transformers/v5.23.0/ko/model_doc/kimi_linear#transformers.KimiLinearModel) (KimiLinearConfig model) - [Kimi_K25Config](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25Config) configuration class: [Kimi_K25Model](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25Model) (Kimi_K25Config model) - [Kimi_K25VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25VisionConfig) configuration class: [Kimi_K25VisionModel](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25VisionModel) (Kimi_K25VisionConfig model) - [Kolibri1Config](/docs/transformers/v5.23.0/ko/model_doc/kolibri1#transformers.Kolibri1Config) configuration class: [Kolibri1Model](/docs/transformers/v5.23.0/ko/model_doc/kolibri1#transformers.Kolibri1Model) (Kolibri1Config model) - [Kosmos2Config](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2Config) configuration class: [Kosmos2Model](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2Model) (Kosmos2Config model) - [Kosmos2_5Config](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5Config) configuration class: [Kosmos2_5Model](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5Model) (Kosmos2_5Config model) - [KyutaiSpeechToTextConfig](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextConfig) configuration class: [KyutaiSpeechToTextModel](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextModel) (KyutaiSpeechToTextConfig model) - [LEDConfig](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.LEDConfig) configuration class: [LEDModel](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.LEDModel) (LEDConfig model) - [LagunaConfig](/docs/transformers/v5.23.0/ko/model_doc/laguna#transformers.LagunaConfig) configuration class: [LagunaModel](/docs/transformers/v5.23.0/ko/model_doc/laguna#transformers.LagunaModel) (LagunaConfig model) - [LasrCTCConfig](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrCTCConfig) configuration class: [LasrForCTC](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrForCTC) (LasrCTCConfig model) - [LasrEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrEncoderConfig) configuration class: [LasrEncoder](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrEncoder) (LasrEncoderConfig model) - [LayoutLMConfig](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMConfig) configuration class: [LayoutLMModel](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMModel) (LayoutLMConfig model) - [LayoutLMv2Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2Config) configuration class: [LayoutLMv2Model](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2Model) (LayoutLMv2Config model) - [LayoutLMv3Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3Config) configuration class: [LayoutLMv3Model](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3Model) (LayoutLMv3Config model) - [LevitConfig](/docs/transformers/v5.23.0/ko/model_doc/levit#transformers.LevitConfig) configuration class: [LevitModel](/docs/transformers/v5.23.0/ko/model_doc/levit#transformers.LevitModel) (LevitConfig model) - [Lfm2Config](/docs/transformers/v5.23.0/ko/model_doc/lfm2#transformers.Lfm2Config) configuration class: [Lfm2Model](/docs/transformers/v5.23.0/ko/model_doc/lfm2#transformers.Lfm2Model) (Lfm2Config model) - [Lfm2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm2_moe#transformers.Lfm2MoeConfig) configuration class: [Lfm2MoeModel](/docs/transformers/v5.23.0/ko/model_doc/lfm2_moe#transformers.Lfm2MoeModel) (Lfm2MoeConfig model) - [Lfm2VlConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm2_vl#transformers.Lfm2VlConfig) configuration class: [Lfm2VlModel](/docs/transformers/v5.23.0/ko/model_doc/lfm2_vl#transformers.Lfm2VlModel) (Lfm2VlConfig model) - [Lfm3Config](/docs/transformers/v5.23.0/ko/model_doc/lfm3#transformers.Lfm3Config) configuration class: [Lfm3Model](/docs/transformers/v5.23.0/ko/model_doc/lfm3#transformers.Lfm3Model) (Lfm3Config model) - [Lfm3VlConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlConfig) configuration class: [Lfm3VlModel](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlModel) (Lfm3VlConfig model) - [Lfm3VlVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlVisionConfig) configuration class: [Lfm3VlVisionModel](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlVisionModel) (Lfm3VlVisionConfig model) - [LightGlueConfig](/docs/transformers/v5.23.0/ko/model_doc/lightglue#transformers.LightGlueConfig) configuration class: [LightGlueForKeypointMatching](/docs/transformers/v5.23.0/ko/model_doc/lightglue#transformers.LightGlueForKeypointMatching) (LightGlueConfig model) - [LightOnOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/lighton_ocr#transformers.LightOnOcrConfig) configuration class: [LightOnOcrModel](/docs/transformers/v5.23.0/ko/model_doc/lighton_ocr#transformers.LightOnOcrModel) (LightOnOcrConfig model) - [LiltConfig](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltConfig) configuration class: [LiltModel](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltModel) (LiltConfig model) - [Llama4Config](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4Config) configuration class: [Llama4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4ForConditionalGeneration) (Llama4Config model) - [Llama4TextConfig](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4TextConfig) configuration class: [Llama4TextModel](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4TextModel) (Llama4TextConfig model) - [LlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaConfig) configuration class: [LlamaModel](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaModel) (LlamaConfig model) - [LlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaConfig) configuration class: [LlavaModel](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaModel) (LlavaConfig model) - [LlavaNextConfig](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextConfig) configuration class: [LlavaNextModel](/docs/transformers/v5.23.0/ko/model_doc/llava_next#transformers.LlavaNextModel) (LlavaNextConfig model) - [LlavaNextVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoConfig) configuration class: [LlavaNextVideoModel](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoModel) (LlavaNextVideoConfig model) - [LlavaOnevisionConfig](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionConfig) configuration class: [LlavaOnevisionModel](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionModel) (LlavaOnevisionConfig model) - [LongT5Config](/docs/transformers/v5.23.0/ko/model_doc/longt5#transformers.LongT5Config) configuration class: [LongT5Model](/docs/transformers/v5.23.0/ko/model_doc/longt5#transformers.LongT5Model) (LongT5Config model) - [LongcatFlashConfig](/docs/transformers/v5.23.0/ko/model_doc/longcat_flash#transformers.LongcatFlashConfig) configuration class: [LongcatFlashModel](/docs/transformers/v5.23.0/ko/model_doc/longcat_flash#transformers.LongcatFlashModel) (LongcatFlashConfig model) - [LongformerConfig](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerConfig) configuration class: [LongformerModel](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerModel) (LongformerConfig model) - [LukeConfig](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeConfig) configuration class: [LukeModel](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeModel) (LukeConfig model) - [LwDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/lw_detr#transformers.LwDetrConfig) configuration class: [LwDetrModel](/docs/transformers/v5.23.0/ko/model_doc/lw_detr#transformers.LwDetrModel) (LwDetrConfig model) - [LxmertConfig](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.LxmertConfig) configuration class: [LxmertModel](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.LxmertModel) (LxmertConfig model) - [M2M100Config](/docs/transformers/v5.23.0/ko/model_doc/m2m_100#transformers.M2M100Config) configuration class: [M2M100Model](/docs/transformers/v5.23.0/ko/model_doc/m2m_100#transformers.M2M100Model) (M2M100Config model) - [MBartConfig](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartConfig) configuration class: [MBartModel](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartModel) (MBartConfig model) - [MLCDVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/mlcd#transformers.MLCDVisionConfig) configuration class: [MLCDVisionModel](/docs/transformers/v5.23.0/ko/model_doc/mlcd#transformers.MLCDVisionModel) (MLCDVisionConfig model) - [MMGroundingDinoConfig](/docs/transformers/v5.23.0/ko/model_doc/mm-grounding-dino#transformers.MMGroundingDinoConfig) configuration class: [MMGroundingDinoModel](/docs/transformers/v5.23.0/ko/model_doc/mm-grounding-dino#transformers.MMGroundingDinoModel) (MMGroundingDinoConfig model) - [MPNetConfig](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetConfig) configuration class: [MPNetModel](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetModel) (MPNetConfig model) - [MT5Config](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5Config) configuration class: [MT5Model](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5Model) (MT5Config model) - [Mamba2Config](/docs/transformers/v5.23.0/ko/model_doc/mamba2#transformers.Mamba2Config) configuration class: [Mamba2Model](/docs/transformers/v5.23.0/ko/model_doc/mamba2#transformers.Mamba2Model) (Mamba2Config model) - [MambaConfig](/docs/transformers/v5.23.0/ko/model_doc/mamba#transformers.MambaConfig) configuration class: [MambaModel](/docs/transformers/v5.23.0/ko/model_doc/mamba#transformers.MambaModel) (MambaConfig model) - [MarianConfig](/docs/transformers/v5.23.0/ko/model_doc/marian#transformers.MarianConfig) configuration class: [MarianModel](/docs/transformers/v5.23.0/ko/model_doc/marian#transformers.MarianModel) (MarianConfig model) - [MarkupLMConfig](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMConfig) configuration class: [MarkupLMModel](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMModel) (MarkupLMConfig model) - [Mask2FormerConfig](/docs/transformers/v5.23.0/ko/model_doc/mask2former#transformers.Mask2FormerConfig) configuration class: [Mask2FormerModel](/docs/transformers/v5.23.0/ko/model_doc/mask2former#transformers.Mask2FormerModel) (Mask2FormerConfig model) - [MaskFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/maskformer#transformers.MaskFormerConfig) configuration class: [MaskFormerModel](/docs/transformers/v5.23.0/ko/model_doc/maskformer#transformers.MaskFormerModel) (MaskFormerConfig model) - `MaskFormerSwinConfig` configuration class: `MaskFormerSwinModel` (MaskFormerSwinConfig model) - [MegatronBertConfig](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertConfig) configuration class: [MegatronBertModel](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertModel) (MegatronBertConfig model) - [MellumConfig](/docs/transformers/v5.23.0/ko/model_doc/mellum#transformers.MellumConfig) configuration class: [MellumModel](/docs/transformers/v5.23.0/ko/model_doc/mellum#transformers.MellumModel) (MellumConfig model) - [MetaClip2Config](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2Config) configuration class: [MetaClip2Model](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2Model) (MetaClip2Config model) - [MgpstrConfig](/docs/transformers/v5.23.0/ko/model_doc/mgp-str#transformers.MgpstrConfig) configuration class: [MgpstrForSceneTextRecognition](/docs/transformers/v5.23.0/ko/model_doc/mgp-str#transformers.MgpstrForSceneTextRecognition) (MgpstrConfig model) - [MiMoV2FlashConfig](/docs/transformers/v5.23.0/ko/model_doc/mimo_v2_flash#transformers.MiMoV2FlashConfig) configuration class: [MiMoV2FlashModel](/docs/transformers/v5.23.0/ko/model_doc/mimo_v2_flash#transformers.MiMoV2FlashModel) (MiMoV2FlashConfig model) - [MimiConfig](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiConfig) configuration class: [MimiModel](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiModel) (MimiConfig model) - [MimiDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiDecoderConfig) configuration class: [MimiDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiDecoderModel) (MimiDecoderConfig model) - [MimiEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiEncoderConfig) configuration class: [MimiEncoderModel](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiEncoderModel) (MimiEncoderConfig model) - [MiniCPM3Config](/docs/transformers/v5.23.0/ko/model_doc/minicpm3#transformers.MiniCPM3Config) configuration class: [MiniCPM3Model](/docs/transformers/v5.23.0/ko/model_doc/minicpm3#transformers.MiniCPM3Model) (MiniCPM3Config model) - [MiniCPMOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrConfig) configuration class: [MiniCPMOcrModel](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrModel) (MiniCPMOcrConfig model) - [MiniCPMOcrVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrVisionConfig) configuration class: [MiniCPMOcrVisionModel](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrVisionModel) (MiniCPMOcrVisionConfig model) - [MiniCPMV4_6Config](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6Config) configuration class: [MiniCPMV4_6Model](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6Model) (MiniCPMV4_6Config model) - [MiniCPMV4_7Config](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7Config) configuration class: [MiniCPMV4_7Model](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7Model) (MiniCPMV4_7Config model) - [MiniCPMV4_7VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7VisionConfig) configuration class: [MiniCPMV4_7VisionModel](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7VisionModel) (MiniCPMV4_7VisionConfig model) - [MiniMaxConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxConfig) configuration class: [MiniMaxModel](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxModel) (MiniMaxConfig model) - [MiniMaxM2Config](/docs/transformers/v5.23.0/ko/model_doc/minimax_m2#transformers.MiniMaxM2Config) configuration class: [MiniMaxM2Model](/docs/transformers/v5.23.0/ko/model_doc/minimax_m2#transformers.MiniMaxM2Model) (MiniMaxM2Config model) - [MiniMaxM3VLConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLConfig) configuration class: [MiniMaxM3VLModel](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLModel) (MiniMaxM3VLConfig model) - [MiniMaxM3VLTextConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLTextConfig) configuration class: [MiniMaxM3VLTextModel](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLTextModel) (MiniMaxM3VLTextConfig model) - [Ministral3Config](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3Config) configuration class: [Ministral3Model](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3Model) (Ministral3Config model) - [MinistralConfig](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralConfig) configuration class: [MinistralModel](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralModel) (MinistralConfig model) - [Mistral3Config](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3Config) configuration class: [Mistral3Model](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3Model) (Mistral3Config model) - [Mistral4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4Config) configuration class: [Mistral4Model](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4Model) (Mistral4Config model) - [MistralConfig](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralConfig) configuration class: [MistralModel](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralModel) (MistralConfig model) - [MistralLarge4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4Config) configuration class: [MistralLarge4Model](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4Model) (MistralLarge4Config model) - [MistralLarge4TextConfig](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4TextConfig) configuration class: [MistralLarge4TextModel](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4TextModel) (MistralLarge4TextConfig model) - [MixtralConfig](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralConfig) configuration class: [MixtralModel](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralModel) (MixtralConfig model) - [MllamaConfig](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaConfig) configuration class: [MllamaModel](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaModel) (MllamaConfig model) - [MobileBertConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertConfig) configuration class: [MobileBertModel](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertModel) (MobileBertConfig model) - [MobileNetV1Config](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v1#transformers.MobileNetV1Config) configuration class: [MobileNetV1Model](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v1#transformers.MobileNetV1Model) (MobileNetV1Config model) - [MobileNetV2Config](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v2#transformers.MobileNetV2Config) configuration class: [MobileNetV2Model](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v2#transformers.MobileNetV2Model) (MobileNetV2Config model) - [MobileViTConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilevit#transformers.MobileViTConfig) configuration class: [MobileViTModel](/docs/transformers/v5.23.0/ko/model_doc/mobilevit#transformers.MobileViTModel) (MobileViTConfig model) - [MobileViTV2Config](/docs/transformers/v5.23.0/ko/model_doc/mobilevitv2#transformers.MobileViTV2Config) configuration class: [MobileViTV2Model](/docs/transformers/v5.23.0/ko/model_doc/mobilevitv2#transformers.MobileViTV2Model) (MobileViTV2Config model) - [ModernBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertConfig) configuration class: [ModernBertModel](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertModel) (ModernBertConfig model) - [ModernBertDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/modernbert-decoder#transformers.ModernBertDecoderConfig) configuration class: [ModernBertDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/modernbert-decoder#transformers.ModernBertDecoderModel) (ModernBertDecoderConfig model) - [ModernVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertConfig) configuration class: [ModernVBertModel](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertModel) (ModernVBertConfig model) - [Molmo2Config](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2Config) configuration class: [Molmo2Model](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2Model) (Molmo2Config model) - [Molmo2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2TextConfig) configuration class: [Molmo2TextModel](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2TextModel) (Molmo2TextConfig model) - [MoonshineConfig](/docs/transformers/v5.23.0/ko/model_doc/moonshine#transformers.MoonshineConfig) configuration class: [MoonshineModel](/docs/transformers/v5.23.0/ko/model_doc/moonshine#transformers.MoonshineModel) (MoonshineConfig model) - [MoonshineStreamingConfig](/docs/transformers/v5.23.0/ko/model_doc/moonshine_streaming#transformers.MoonshineStreamingConfig) configuration class: [MoonshineStreamingModel](/docs/transformers/v5.23.0/ko/model_doc/moonshine_streaming#transformers.MoonshineStreamingModel) (MoonshineStreamingConfig model) - [MoshiConfig](/docs/transformers/v5.23.0/ko/model_doc/moshi#transformers.MoshiConfig) configuration class: [MoshiModel](/docs/transformers/v5.23.0/ko/model_doc/moshi#transformers.MoshiModel) (MoshiConfig model) - [MossTranscribeDiarizeAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeAudioConfig) configuration class: [MossTranscribeDiarizeAudioModel](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeAudioModel) (MossTranscribeDiarizeAudioConfig model) - [MossTranscribeDiarizeConfig](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeConfig) configuration class: [MossTranscribeDiarizeModel](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeModel) (MossTranscribeDiarizeConfig model) - [MptConfig](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptConfig) configuration class: [MptModel](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptModel) (MptConfig model) - [MraConfig](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraConfig) configuration class: [MraModel](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraModel) (MraConfig model) - [MuseGlimmerAssistantConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer_assistant#transformers.MuseGlimmerAssistantConfig) configuration class: [MuseGlimmerAssistantModel](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer_assistant#transformers.MuseGlimmerAssistantModel) (MuseGlimmerAssistantConfig model) - [MuseGlimmerConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerConfig) configuration class: [MuseGlimmerModel](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerModel) (MuseGlimmerConfig model) - [MuseGlimmerTextConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerTextConfig) configuration class: [MuseGlimmerTextModel](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerTextModel) (MuseGlimmerTextConfig model) - [MuseGlimmerVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerVisionConfig) configuration class: [MuseGlimmerVisionModel](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerVisionModel) (MuseGlimmerVisionConfig model) - [MuseSparkAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkAudioConfig) configuration class: [MuseSparkAudioModel](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkAudioModel) (MuseSparkAudioConfig model) - [MuseSparkConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkConfig) configuration class: [MuseSparkModel](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkModel) (MuseSparkConfig model) - [MuseSparkTextConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkTextConfig) configuration class: [MuseSparkTextModel](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkTextModel) (MuseSparkTextConfig model) - [MuseSparkVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkVisionConfig) configuration class: [MuseSparkVisionModel](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkVisionModel) (MuseSparkVisionConfig model) - [MusicFlamingoConfig](/docs/transformers/v5.23.0/ko/model_doc/musicflamingo#transformers.MusicFlamingoConfig) configuration class: [MusicFlamingoModel](/docs/transformers/v5.23.0/ko/model_doc/musicflamingo#transformers.MusicFlamingoModel) (MusicFlamingoConfig model) - [MusicgenConfig](/docs/transformers/v5.23.0/ko/model_doc/musicgen#transformers.MusicgenConfig) configuration class: [MusicgenModel](/docs/transformers/v5.23.0/ko/model_doc/musicgen#transformers.MusicgenModel) (MusicgenConfig model) - [MusicgenMelodyConfig](/docs/transformers/v5.23.0/ko/model_doc/musicgen_melody#transformers.MusicgenMelodyConfig) configuration class: [MusicgenMelodyModel](/docs/transformers/v5.23.0/ko/model_doc/musicgen_melody#transformers.MusicgenMelodyModel) (MusicgenMelodyConfig model) - [MvpConfig](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpConfig) configuration class: [MvpModel](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpModel) (MvpConfig model) - [NanoChatConfig](/docs/transformers/v5.23.0/ko/model_doc/nanochat#transformers.NanoChatConfig) configuration class: [NanoChatModel](/docs/transformers/v5.23.0/ko/model_doc/nanochat#transformers.NanoChatModel) (NanoChatConfig model) - [Nemotron3DiarizationAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationAudioConfig) configuration class: [Nemotron3DiarizationAudioModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationAudioModel) (Nemotron3DiarizationAudioConfig model) - [Nemotron3DiarizationConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationConfig) configuration class: [Nemotron3DiarizationModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationModel) (Nemotron3DiarizationConfig model) - [Nemotron3_5AsrConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_asr#transformers.Nemotron3_5AsrConfig) configuration class: [Nemotron3_5AsrForRNNT](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_asr#transformers.Nemotron3_5AsrForRNNT) (Nemotron3_5AsrConfig model) - [Nemotron3_5TranscribeAsrAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeAsrAudioConfig) configuration class: [Nemotron3_5TranscribeAsrAudioModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeAsrAudioModel) (Nemotron3_5TranscribeAsrAudioConfig model) - [Nemotron3_5TranscribeAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeAudioConfig) configuration class: [Nemotron3_5TranscribeAudioModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeAudioModel) (Nemotron3_5TranscribeAudioConfig model) - [Nemotron3_5TranscribeConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeConfig) configuration class: [Nemotron3_5TranscribeModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeModel) (Nemotron3_5TranscribeConfig model) - [Nemotron3_5TranscribeWordTimestampConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeWordTimestampConfig) configuration class: [Nemotron3_5TranscribeWordTimestampModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeWordTimestampModel) (Nemotron3_5TranscribeWordTimestampConfig model) - [NemotronAsrStreamingConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingConfig) configuration class: [NemotronAsrStreamingForRNNT](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingForRNNT) (NemotronAsrStreamingConfig model) - [NemotronAsrStreamingEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingEncoderConfig) configuration class: [NemotronAsrStreamingEncoder](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingEncoder) (NemotronAsrStreamingEncoderConfig model) - [NemotronConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronConfig) configuration class: [NemotronModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronModel) (NemotronConfig model) - [NemotronHConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h#transformers.NemotronHConfig) configuration class: [NemotronHModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h#transformers.NemotronHModel) (NemotronHConfig model) - [NeoMMEConfig](/docs/transformers/v5.23.0/ko/model_doc/neomme#transformers.NeoMMEConfig) configuration class: [NeoMMEModel](/docs/transformers/v5.23.0/ko/model_doc/neomme#transformers.NeoMMEModel) (NeoMMEConfig model) - [NeuCodecConfig](/docs/transformers/v5.23.0/ko/model_doc/neucodec#transformers.NeuCodecConfig) configuration class: [NeuCodecModel](/docs/transformers/v5.23.0/ko/model_doc/neucodec#transformers.NeuCodecModel) (NeuCodecConfig model) - [NllbMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/nllb-moe#transformers.NllbMoeConfig) configuration class: [NllbMoeModel](/docs/transformers/v5.23.0/ko/model_doc/nllb-moe#transformers.NllbMoeModel) (NllbMoeConfig model) - [NomicBertConfig](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertConfig) configuration class: [NomicBertModel](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertModel) (NomicBertConfig model) - [NystromformerConfig](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerConfig) configuration class: [NystromformerModel](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerModel) (NystromformerConfig model) - [OPTConfig](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTConfig) configuration class: [OPTModel](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTModel) (OPTConfig model) - [Olmo2Config](/docs/transformers/v5.23.0/ko/model_doc/olmo2#transformers.Olmo2Config) configuration class: [Olmo2Model](/docs/transformers/v5.23.0/ko/model_doc/olmo2#transformers.Olmo2Model) (Olmo2Config model) - [Olmo3Config](/docs/transformers/v5.23.0/ko/model_doc/olmo3#transformers.Olmo3Config) configuration class: [Olmo3Model](/docs/transformers/v5.23.0/ko/model_doc/olmo3#transformers.Olmo3Model) (Olmo3Config model) - [OlmoConfig](/docs/transformers/v5.23.0/ko/model_doc/olmo#transformers.OlmoConfig) configuration class: [OlmoModel](/docs/transformers/v5.23.0/ko/model_doc/olmo#transformers.OlmoModel) (OlmoConfig model) - [OlmoHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/olmo_hybrid#transformers.OlmoHybridConfig) configuration class: [OlmoHybridModel](/docs/transformers/v5.23.0/ko/model_doc/olmo_hybrid#transformers.OlmoHybridModel) (OlmoHybridConfig model) - [OlmoeConfig](/docs/transformers/v5.23.0/ko/model_doc/olmoe#transformers.OlmoeConfig) configuration class: [OlmoeModel](/docs/transformers/v5.23.0/ko/model_doc/olmoe#transformers.OlmoeModel) (OlmoeConfig model) - [OmDetTurboConfig](/docs/transformers/v5.23.0/ko/model_doc/omdet-turbo#transformers.OmDetTurboConfig) configuration class: [OmDetTurboForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/omdet-turbo#transformers.OmDetTurboForObjectDetection) (OmDetTurboConfig model) - [OneFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/oneformer#transformers.OneFormerConfig) configuration class: [OneFormerModel](/docs/transformers/v5.23.0/ko/model_doc/oneformer#transformers.OneFormerModel) (OneFormerConfig model) - [OpenAIGPTConfig](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTConfig) configuration class: [OpenAIGPTModel](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTModel) (OpenAIGPTConfig model) - [OpenAIPrivacyFilterConfig](/docs/transformers/v5.23.0/ko/model_doc/openai_privacy_filter#transformers.OpenAIPrivacyFilterConfig) configuration class: [OpenAIPrivacyFilterModel](/docs/transformers/v5.23.0/ko/model_doc/openai_privacy_filter#transformers.OpenAIPrivacyFilterModel) (OpenAIPrivacyFilterConfig model) - [Ovis2Config](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2Config) configuration class: [Ovis2Model](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2Model) (Ovis2Config model) - [OwlViTConfig](/docs/transformers/v5.23.0/ko/model_doc/owlvit#transformers.OwlViTConfig) configuration class: [OwlViTModel](/docs/transformers/v5.23.0/ko/model_doc/owlvit#transformers.OwlViTModel) (OwlViTConfig model) - [Owlv2Config](/docs/transformers/v5.23.0/ko/model_doc/owlv2#transformers.Owlv2Config) configuration class: [Owlv2Model](/docs/transformers/v5.23.0/ko/model_doc/owlv2#transformers.Owlv2Model) (Owlv2Config model) - [PI0Config](/docs/transformers/v5.23.0/ko/model_doc/pi0#transformers.PI0Config) configuration class: [PI0Model](/docs/transformers/v5.23.0/ko/model_doc/pi0#transformers.PI0Model) (PI0Config model) - [PLBartConfig](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartConfig) configuration class: [PLBartModel](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartModel) (PLBartConfig model) - [PPDocLayoutV3Config](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v3#transformers.PPDocLayoutV3Config) configuration class: [PPDocLayoutV3Model](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v3#transformers.PPDocLayoutV3Model) (PPDocLayoutV3Config model) - [PPDocLayoutV4Config](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v4#transformers.PPDocLayoutV4Config) configuration class: [PPDocLayoutV4Model](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v4#transformers.PPDocLayoutV4Model) (PPDocLayoutV4Config model) - [PPOCRV5MobileRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_mobile_rec#transformers.PPOCRV5MobileRecConfig) configuration class: [PPOCRV5MobileRecModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_mobile_rec#transformers.PPOCRV5MobileRecModel) (PPOCRV5MobileRecConfig model) - [PPOCRV5ServerRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_server_rec#transformers.PPOCRV5ServerRecConfig) configuration class: [PPOCRV5ServerRecModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_server_rec#transformers.PPOCRV5ServerRecModel) (PPOCRV5ServerRecConfig model) - [PPOCRV6MediumDetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_medium_det#transformers.PPOCRV6MediumDetConfig) configuration class: [PPOCRV6MediumDetModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_medium_det#transformers.PPOCRV6MediumDetModel) (PPOCRV6MediumDetConfig model) - [PPOCRV6SmallDetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_det#transformers.PPOCRV6SmallDetConfig) configuration class: [PPOCRV6SmallDetModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_det#transformers.PPOCRV6SmallDetModel) (PPOCRV6SmallDetConfig model) - [PPOCRV6SmallRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_rec#transformers.PPOCRV6SmallRecConfig) configuration class: [PPOCRV6SmallRecModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_rec#transformers.PPOCRV6SmallRecModel) (PPOCRV6SmallRecConfig model) - [PPOCRV6TinyRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_tiny_rec#transformers.PPOCRV6TinyRecConfig) configuration class: [PPOCRV6TinyRecModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_tiny_rec#transformers.PPOCRV6TinyRecModel) (PPOCRV6TinyRecConfig model) - [PaddleOCRVLConfig](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVLConfig) configuration class: [PaddleOCRVLModel](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVLModel) (PaddleOCRVLConfig model) - [PaliGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaConfig) configuration class: [PaliGemmaModel](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaModel) (PaliGemmaConfig model) - [ParakeetCTCConfig](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetCTCConfig) configuration class: [ParakeetForCTC](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForCTC) (ParakeetCTCConfig model) - [ParakeetEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetEncoderConfig) configuration class: [ParakeetEncoder](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetEncoder) (ParakeetEncoderConfig model) - [ParakeetRNNTConfig](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetRNNTConfig) configuration class: [ParakeetForRNNT](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForRNNT) (ParakeetRNNTConfig model) - [ParakeetTDTConfig](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetTDTConfig) configuration class: [ParakeetForTDT](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForTDT) (ParakeetTDTConfig model) - [PatchTSMixerConfig](/docs/transformers/v5.23.0/ko/model_doc/patchtsmixer#transformers.PatchTSMixerConfig) configuration class: [PatchTSMixerModel](/docs/transformers/v5.23.0/ko/model_doc/patchtsmixer#transformers.PatchTSMixerModel) (PatchTSMixerConfig model) - [PatchTSTConfig](/docs/transformers/v5.23.0/ko/model_doc/patchtst#transformers.PatchTSTConfig) configuration class: [PatchTSTModel](/docs/transformers/v5.23.0/ko/model_doc/patchtst#transformers.PatchTSTModel) (PatchTSTConfig model) - [PeAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_audio#transformers.PeAudioConfig) configuration class: [PeAudioModel](/docs/transformers/v5.23.0/ko/model_doc/pe_audio#transformers.PeAudioModel) (PeAudioConfig model) - [PeAudioEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_audio#transformers.PeAudioEncoderConfig) configuration class: [PeAudioEncoder](/docs/transformers/v5.23.0/ko/model_doc/pe_audio#transformers.PeAudioEncoder) (PeAudioEncoderConfig model) - [PeAudioVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_audio_video#transformers.PeAudioVideoConfig) configuration class: [PeAudioVideoModel](/docs/transformers/v5.23.0/ko/model_doc/pe_audio_video#transformers.PeAudioVideoModel) (PeAudioVideoConfig model) - [PeAudioVideoEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_audio_video#transformers.PeAudioVideoEncoderConfig) configuration class: [PeAudioVideoEncoder](/docs/transformers/v5.23.0/ko/model_doc/pe_audio_video#transformers.PeAudioVideoEncoder) (PeAudioVideoEncoderConfig model) - [PeVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_video#transformers.PeVideoConfig) configuration class: [PeVideoModel](/docs/transformers/v5.23.0/ko/model_doc/pe_video#transformers.PeVideoModel) (PeVideoConfig model) - [PeVideoEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/pe_video#transformers.PeVideoEncoderConfig) configuration class: [PeVideoEncoder](/docs/transformers/v5.23.0/ko/model_doc/pe_video#transformers.PeVideoEncoder) (PeVideoEncoderConfig model) - [PegasusConfig](/docs/transformers/v5.23.0/ko/model_doc/pegasus#transformers.PegasusConfig) configuration class: [PegasusModel](/docs/transformers/v5.23.0/ko/model_doc/pegasus#transformers.PegasusModel) (PegasusConfig model) - [PegasusXConfig](/docs/transformers/v5.23.0/ko/model_doc/pegasus_x#transformers.PegasusXConfig) configuration class: [PegasusXModel](/docs/transformers/v5.23.0/ko/model_doc/pegasus_x#transformers.PegasusXModel) (PegasusXConfig model) - [PerceiverConfig](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverConfig) configuration class: [PerceiverModel](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverModel) (PerceiverConfig model) - [PerceptionLMConfig](/docs/transformers/v5.23.0/ko/model_doc/perception_lm#transformers.PerceptionLMConfig) configuration class: [PerceptionLMModel](/docs/transformers/v5.23.0/ko/model_doc/perception_lm#transformers.PerceptionLMModel) (PerceptionLMConfig model) - [PersimmonConfig](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonConfig) configuration class: [PersimmonModel](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonModel) (PersimmonConfig model) - [Phi3Config](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3Config) configuration class: [Phi3Model](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3Model) (Phi3Config model) - [Phi4MultimodalConfig](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalConfig) configuration class: [Phi4MultimodalModel](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalModel) (Phi4MultimodalConfig model) - [PhiConfig](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiConfig) configuration class: [PhiModel](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiModel) (PhiConfig model) - [PhimoeConfig](/docs/transformers/v5.23.0/ko/model_doc/phimoe#transformers.PhimoeConfig) configuration class: [PhimoeModel](/docs/transformers/v5.23.0/ko/model_doc/phimoe#transformers.PhimoeModel) (PhimoeConfig model) - [PixioConfig](/docs/transformers/v5.23.0/ko/model_doc/pixio#transformers.PixioConfig) configuration class: [PixioModel](/docs/transformers/v5.23.0/ko/model_doc/pixio#transformers.PixioModel) (PixioConfig model) - [PixtralVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/pixtral#transformers.PixtralVisionConfig) configuration class: [PixtralVisionModel](/docs/transformers/v5.23.0/ko/model_doc/pixtral#transformers.PixtralVisionModel) (PixtralVisionConfig model) - [PocketTTSAudioTokenizerConfig](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerConfig) configuration class: [PocketTTSAudioTokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerModel) (PocketTTSAudioTokenizerConfig model) - [PocketTTSAudioTokenizerDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerDecoderConfig) configuration class: [PocketTTSAudioTokenizerDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerDecoderModel) (PocketTTSAudioTokenizerDecoderConfig model) - [PocketTTSAudioTokenizerEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerEncoderConfig) configuration class: [PocketTTSAudioTokenizerEncoderModel](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerEncoderModel) (PocketTTSAudioTokenizerEncoderConfig model) - [PocketTTSConfig](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts#transformers.PocketTTSConfig) configuration class: [PocketTTSModel](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts#transformers.PocketTTSModel) (PocketTTSConfig model) - [PoolFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/poolformer#transformers.PoolFormerConfig) configuration class: [PoolFormerModel](/docs/transformers/v5.23.0/ko/model_doc/poolformer#transformers.PoolFormerModel) (PoolFormerConfig model) - [ProphetNetConfig](/docs/transformers/v5.23.0/ko/model_doc/prophetnet#transformers.ProphetNetConfig) configuration class: [ProphetNetModel](/docs/transformers/v5.23.0/ko/model_doc/prophetnet#transformers.ProphetNetModel) (ProphetNetConfig model) - [PvtConfig](/docs/transformers/v5.23.0/ko/model_doc/pvt#transformers.PvtConfig) configuration class: [PvtModel](/docs/transformers/v5.23.0/ko/model_doc/pvt#transformers.PvtModel) (PvtConfig model) - [PvtV2Config](/docs/transformers/v5.23.0/ko/model_doc/pvt_v2#transformers.PvtV2Config) configuration class: [PvtV2Model](/docs/transformers/v5.23.0/ko/model_doc/pvt_v2#transformers.PvtV2Model) (PvtV2Config model) - [QianfanOCRConfig](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRConfig) configuration class: [QianfanOCRModel](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRModel) (QianfanOCRConfig model) - [QianfanOCRVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRVisionConfig) configuration class: [QianfanOCRVisionModel](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRVisionModel) (QianfanOCRVisionConfig model) - [Qwen2AudioConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioConfig) configuration class: [Qwen2AudioModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioModel) (Qwen2AudioConfig model) - [Qwen2AudioEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioEncoderConfig) configuration class: [Qwen2AudioEncoder](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioEncoder) (Qwen2AudioEncoderConfig model) - [Qwen2Config](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Config) configuration class: [Qwen2Model](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Model) (Qwen2Config model) - [Qwen2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeConfig) configuration class: [Qwen2MoeModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeModel) (Qwen2MoeConfig model) - [Qwen2VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLConfig) configuration class: [Qwen2VLModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLModel) (Qwen2VLConfig model) - [Qwen2VLTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLTextConfig) configuration class: [Qwen2VLTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLTextModel) (Qwen2VLTextConfig model) - [Qwen2_5OmniThinkerConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniThinkerConfig) configuration class: [Qwen2_5OmniThinkerModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniThinkerModel) (Qwen2_5OmniThinkerConfig model) - [Qwen2_5_VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLConfig) configuration class: [Qwen2_5_VLModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLModel) (Qwen2_5_VLConfig model) - [Qwen2_5_VLTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLTextConfig) configuration class: [Qwen2_5_VLTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLTextModel) (Qwen2_5_VLTextConfig model) - [Qwen2_5_VLVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLVisionConfig) configuration class: [Qwen2_5_VisionTransformerPretrainedModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VisionTransformerPretrainedModel) (Qwen2_5_VLVisionConfig model) - [Qwen3ASRConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRConfig) configuration class: [Qwen3ASRModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRModel) (Qwen3ASRConfig model) - [Qwen3ASREncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASREncoderConfig) configuration class: [Qwen3ASREncoder](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASREncoder) (Qwen3ASREncoderConfig model) - [Qwen3Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3Config) configuration class: [Qwen3Model](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3Model) (Qwen3Config model) - [Qwen3MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeConfig) configuration class: [Qwen3MoeModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeModel) (Qwen3MoeConfig model) - [Qwen3NextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextConfig) configuration class: [Qwen3NextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextModel) (Qwen3NextConfig model) - [Qwen3OmniMoeThinkerConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeThinkerConfig) configuration class: [Qwen3OmniMoeThinkerModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeThinkerModel) (Qwen3OmniMoeThinkerConfig model) - [Qwen3VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLConfig) configuration class: [Qwen3VLModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLModel) (Qwen3VLConfig model) - [Qwen3VLMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeConfig) configuration class: [Qwen3VLMoeModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeModel) (Qwen3VLMoeConfig model) - [Qwen3VLMoeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeTextConfig) configuration class: [Qwen3VLMoeTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeTextModel) (Qwen3VLMoeTextConfig model) - [Qwen3VLMoeVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeVisionConfig) configuration class: [Qwen3VLMoeVisionModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeVisionModel) (Qwen3VLMoeVisionConfig model) - [Qwen3VLTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLTextConfig) configuration class: [Qwen3VLTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLTextModel) (Qwen3VLTextConfig model) - [Qwen3VLVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLVisionConfig) configuration class: [Qwen3VLVisionModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLVisionModel) (Qwen3VLVisionConfig model) - [Qwen3_5Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Config) configuration class: [Qwen3_5Model](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Model) (Qwen3_5Config model) - [Qwen3_5MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeConfig) configuration class: [Qwen3_5MoeModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeModel) (Qwen3_5MoeConfig model) - [Qwen3_5MoeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeTextConfig) configuration class: [Qwen3_5MoeTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeTextModel) (Qwen3_5MoeTextConfig model) - [Qwen3_5MoeVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeVisionConfig) configuration class: [Qwen3_5MoeVisionModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeVisionModel) (Qwen3_5MoeVisionConfig model) - [Qwen3_5TextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5TextConfig) configuration class: [Qwen3_5TextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5TextModel) (Qwen3_5TextConfig model) - [Qwen3_5VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5VisionConfig) configuration class: [Qwen3_5VisionModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5VisionModel) (Qwen3_5VisionConfig model) - [Qwen4ExpConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpConfig) configuration class: [Qwen4ExpModel](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpModel) (Qwen4ExpConfig model) - [Qwen4ExpTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpTextConfig) configuration class: [Qwen4ExpTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpTextModel) (Qwen4ExpTextConfig model) - [Qwen4ExpVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpVisionConfig) configuration class: [Qwen4ExpVisionModel](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpVisionModel) (Qwen4ExpVisionConfig model) - [RTDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/rt_detr#transformers.RTDetrConfig) configuration class: [RTDetrModel](/docs/transformers/v5.23.0/ko/model_doc/rt_detr#transformers.RTDetrModel) (RTDetrConfig model) - [RTDetrV2Config](/docs/transformers/v5.23.0/ko/model_doc/rt_detr_v2#transformers.RTDetrV2Config) configuration class: [RTDetrV2Model](/docs/transformers/v5.23.0/ko/model_doc/rt_detr_v2#transformers.RTDetrV2Model) (RTDetrV2Config model) - [RadioConfig](/docs/transformers/v5.23.0/ko/model_doc/radio#transformers.RadioConfig) configuration class: [RadioModel](/docs/transformers/v5.23.0/ko/model_doc/radio#transformers.RadioModel) (RadioConfig model) - [RecurrentGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/recurrent_gemma#transformers.RecurrentGemmaConfig) configuration class: [RecurrentGemmaModel](/docs/transformers/v5.23.0/ko/model_doc/recurrent_gemma#transformers.RecurrentGemmaModel) (RecurrentGemmaConfig model) - [ReformerConfig](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerConfig) configuration class: [ReformerModel](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerModel) (ReformerConfig model) - [RegNetConfig](/docs/transformers/v5.23.0/ko/model_doc/regnet#transformers.RegNetConfig) configuration class: [RegNetModel](/docs/transformers/v5.23.0/ko/model_doc/regnet#transformers.RegNetModel) (RegNetConfig model) - [RemBertConfig](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertConfig) configuration class: [RemBertModel](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertModel) (RemBertConfig model) - [ResNetConfig](/docs/transformers/v5.23.0/ko/model_doc/resnet#transformers.ResNetConfig) configuration class: [ResNetModel](/docs/transformers/v5.23.0/ko/model_doc/resnet#transformers.ResNetModel) (ResNetConfig model) - [RfDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/rf_detr#transformers.RfDetrConfig) configuration class: [RfDetrModel](/docs/transformers/v5.23.0/ko/model_doc/rf_detr#transformers.RfDetrModel) (RfDetrConfig model) - [RoCBertConfig](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertConfig) configuration class: [RoCBertModel](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertModel) (RoCBertConfig model) - [RoFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerConfig) configuration class: [RoFormerModel](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerModel) (RoFormerConfig model) - [RobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaConfig) configuration class: [RobertaModel](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaModel) (RobertaConfig model) - [RobertaPreLayerNormConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormConfig) configuration class: [RobertaPreLayerNormModel](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormModel) (RobertaPreLayerNormConfig model) - [RwkvConfig](/docs/transformers/v5.23.0/ko/model_doc/rwkv#transformers.RwkvConfig) configuration class: [RwkvModel](/docs/transformers/v5.23.0/ko/model_doc/rwkv#transformers.RwkvModel) (RwkvConfig model) - [SEWConfig](/docs/transformers/v5.23.0/ko/model_doc/sew#transformers.SEWConfig) configuration class: [SEWModel](/docs/transformers/v5.23.0/ko/model_doc/sew#transformers.SEWModel) (SEWConfig model) - [SEWDConfig](/docs/transformers/v5.23.0/ko/model_doc/sew-d#transformers.SEWDConfig) configuration class: [SEWDModel](/docs/transformers/v5.23.0/ko/model_doc/sew-d#transformers.SEWDModel) (SEWDConfig model) - [Sam2Config](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2Config) configuration class: [Sam2Model](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2Model) (Sam2Config model) - [Sam2HieraDetConfig](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2HieraDetConfig) configuration class: [Sam2HieraDetModel](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2HieraDetModel) (Sam2HieraDetConfig model) - [Sam2VideoConfig](/docs/transformers/v5.23.0/ko/model_doc/sam2_video#transformers.Sam2VideoConfig) configuration class: [Sam2VideoModel](/docs/transformers/v5.23.0/ko/model_doc/sam2_video#transformers.Sam2VideoModel) (Sam2VideoConfig model) - [Sam2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2VisionConfig) configuration class: [Sam2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2VisionModel) (Sam2VisionConfig model) - [Sam3Config](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3Config) configuration class: [Sam3Model](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3Model) (Sam3Config model) - [Sam3LiteTextConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextConfig) configuration class: [Sam3LiteTextModel](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextModel) (Sam3LiteTextConfig model) - [Sam3LiteTextTextConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextTextConfig) configuration class: [Sam3LiteTextTextModel](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextTextModel) (Sam3LiteTextTextConfig model) - [Sam3TrackerConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker#transformers.Sam3TrackerConfig) configuration class: [Sam3TrackerModel](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker#transformers.Sam3TrackerModel) (Sam3TrackerConfig model) - [Sam3TrackerVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker_video#transformers.Sam3TrackerVideoConfig) configuration class: [Sam3TrackerVideoModel](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker_video#transformers.Sam3TrackerVideoModel) (Sam3TrackerVideoConfig model) - [Sam3ViTConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3ViTConfig) configuration class: [Sam3ViTModel](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3ViTModel) (Sam3ViTConfig model) - [Sam3VideoConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3_video#transformers.Sam3VideoConfig) configuration class: [Sam3VideoModel](/docs/transformers/v5.23.0/ko/model_doc/sam3_video#transformers.Sam3VideoModel) (Sam3VideoConfig model) - [Sam3VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3VisionConfig) configuration class: [Sam3VisionModel](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3VisionModel) (Sam3VisionConfig model) - [SamConfig](/docs/transformers/v5.23.0/ko/model_doc/sam#transformers.SamConfig) configuration class: [SamModel](/docs/transformers/v5.23.0/ko/model_doc/sam#transformers.SamModel) (SamConfig model) - [SamHQConfig](/docs/transformers/v5.23.0/ko/model_doc/sam_hq#transformers.SamHQConfig) configuration class: [SamHQModel](/docs/transformers/v5.23.0/ko/model_doc/sam_hq#transformers.SamHQModel) (SamHQConfig model) - [SamHQVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/sam_hq#transformers.SamHQVisionConfig) configuration class: [SamHQVisionModel](/docs/transformers/v5.23.0/ko/model_doc/sam_hq#transformers.SamHQVisionModel) (SamHQVisionConfig model) - [SamVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/sam#transformers.SamVisionConfig) configuration class: [SamVisionModel](/docs/transformers/v5.23.0/ko/model_doc/sam#transformers.SamVisionModel) (SamVisionConfig model) - [Sapiens2Config](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2Config) configuration class: [Sapiens2Model](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2Model) (Sapiens2Config model) - [SeamlessM4TConfig](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TConfig) configuration class: [SeamlessM4TModel](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TModel) (SeamlessM4TConfig model) - [SeamlessM4Tv2Config](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t_v2#transformers.SeamlessM4Tv2Config) configuration class: [SeamlessM4Tv2Model](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t_v2#transformers.SeamlessM4Tv2Model) (SeamlessM4Tv2Config model) - [SeedOssConfig](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssConfig) configuration class: [SeedOssModel](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssModel) (SeedOssConfig model) - [SegGptConfig](/docs/transformers/v5.23.0/ko/model_doc/seggpt#transformers.SegGptConfig) configuration class: [SegGptModel](/docs/transformers/v5.23.0/ko/model_doc/seggpt#transformers.SegGptModel) (SegGptConfig model) - [SegformerConfig](/docs/transformers/v5.23.0/ko/model_doc/segformer#transformers.SegformerConfig) configuration class: [SegformerModel](/docs/transformers/v5.23.0/ko/model_doc/segformer#transformers.SegformerModel) (SegformerConfig model) - [Siglip2Config](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2Config) configuration class: [Siglip2Model](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2Model) (Siglip2Config model) - [Siglip2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2VisionConfig) configuration class: [Siglip2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2VisionModel) (Siglip2VisionConfig model) - [SiglipConfig](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipConfig) configuration class: [SiglipModel](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipModel) (SiglipConfig model) - [SiglipVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipVisionConfig) configuration class: [SiglipVisionModel](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipVisionModel) (SiglipVisionConfig model) - [SmolLM3Config](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3Config) configuration class: [SmolLM3Model](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3Model) (SmolLM3Config model) - [SmolVLMConfig](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMConfig) configuration class: [SmolVLMModel](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMModel) (SmolVLMConfig model) - [SmolVLMVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMVisionConfig) configuration class: [SmolVLMVisionTransformer](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMVisionTransformer) (SmolVLMVisionConfig model) - [SolarOpenConfig](/docs/transformers/v5.23.0/ko/model_doc/solar_open#transformers.SolarOpenConfig) configuration class: [SolarOpenModel](/docs/transformers/v5.23.0/ko/model_doc/solar_open#transformers.SolarOpenModel) (SolarOpenConfig model) - [Speech2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/speech_to_text#transformers.Speech2TextConfig) configuration class: [Speech2TextModel](/docs/transformers/v5.23.0/ko/model_doc/speech_to_text#transformers.Speech2TextModel) (Speech2TextConfig model) - [SpeechT5Config](/docs/transformers/v5.23.0/ko/model_doc/speecht5#transformers.SpeechT5Config) configuration class: [SpeechT5Model](/docs/transformers/v5.23.0/ko/model_doc/speecht5#transformers.SpeechT5Model) (SpeechT5Config model) - [SplinterConfig](/docs/transformers/v5.23.0/ko/model_doc/splinter#transformers.SplinterConfig) configuration class: [SplinterModel](/docs/transformers/v5.23.0/ko/model_doc/splinter#transformers.SplinterModel) (SplinterConfig model) - [SqueezeBertConfig](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertConfig) configuration class: [SqueezeBertModel](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertModel) (SqueezeBertConfig model) - [StableLmConfig](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmConfig) configuration class: [StableLmModel](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmModel) (StableLmConfig model) - [Starcoder2Config](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2Config) configuration class: [Starcoder2Model](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2Model) (Starcoder2Config model) - [Step3p7Config](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7Config) configuration class: [Step3p7Model](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7Model) (Step3p7Config model) - [SwiftFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/swiftformer#transformers.SwiftFormerConfig) configuration class: [SwiftFormerModel](/docs/transformers/v5.23.0/ko/model_doc/swiftformer#transformers.SwiftFormerModel) (SwiftFormerConfig model) - [Swin2SRConfig](/docs/transformers/v5.23.0/ko/model_doc/swin2sr#transformers.Swin2SRConfig) configuration class: [Swin2SRModel](/docs/transformers/v5.23.0/ko/model_doc/swin2sr#transformers.Swin2SRModel) (Swin2SRConfig model) - [SwinConfig](/docs/transformers/v5.23.0/ko/model_doc/swin#transformers.SwinConfig) configuration class: [SwinModel](/docs/transformers/v5.23.0/ko/model_doc/swin#transformers.SwinModel) (SwinConfig model) - [Swinv2Config](/docs/transformers/v5.23.0/ko/model_doc/swinv2#transformers.Swinv2Config) configuration class: [Swinv2Model](/docs/transformers/v5.23.0/ko/model_doc/swinv2#transformers.Swinv2Model) (Swinv2Config model) - [SwitchTransformersConfig](/docs/transformers/v5.23.0/ko/model_doc/switch_transformers#transformers.SwitchTransformersConfig) configuration class: [SwitchTransformersModel](/docs/transformers/v5.23.0/ko/model_doc/switch_transformers#transformers.SwitchTransformersModel) (SwitchTransformersConfig model) - [T5Config](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Config) configuration class: [T5Model](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Model) (T5Config model) - [T5Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2Config) configuration class: [T5Gemma2Model](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2Model) (T5Gemma2Config model) - [T5Gemma2EncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2EncoderConfig) configuration class: `T5Gemma2Encoder` (T5Gemma2EncoderConfig model) - [T5GemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaConfig) configuration class: [T5GemmaModel](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaModel) (T5GemmaConfig model) - [TableTransformerConfig](/docs/transformers/v5.23.0/ko/model_doc/table-transformer#transformers.TableTransformerConfig) configuration class: [TableTransformerModel](/docs/transformers/v5.23.0/ko/model_doc/table-transformer#transformers.TableTransformerModel) (TableTransformerConfig model) - [TapasConfig](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasConfig) configuration class: [TapasModel](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasModel) (TapasConfig model) - [TextNetConfig](/docs/transformers/v5.23.0/ko/model_doc/textnet#transformers.TextNetConfig) configuration class: [TextNetModel](/docs/transformers/v5.23.0/ko/model_doc/textnet#transformers.TextNetModel) (TextNetConfig model) - [TimeSeriesTransformerConfig](/docs/transformers/v5.23.0/ko/model_doc/time_series_transformer#transformers.TimeSeriesTransformerConfig) configuration class: [TimeSeriesTransformerModel](/docs/transformers/v5.23.0/ko/model_doc/time_series_transformer#transformers.TimeSeriesTransformerModel) (TimeSeriesTransformerConfig model) - [TimesFm2_5Config](/docs/transformers/v5.23.0/ko/model_doc/timesfm2_5#transformers.TimesFm2_5Config) configuration class: [TimesFm2_5Model](/docs/transformers/v5.23.0/ko/model_doc/timesfm2_5#transformers.TimesFm2_5Model) (TimesFm2_5Config model) - [TimesFmConfig](/docs/transformers/v5.23.0/ko/model_doc/timesfm#transformers.TimesFmConfig) configuration class: [TimesFmModel](/docs/transformers/v5.23.0/ko/model_doc/timesfm#transformers.TimesFmModel) (TimesFmConfig model) - [TimesformerConfig](/docs/transformers/v5.23.0/ko/model_doc/timesformer#transformers.TimesformerConfig) configuration class: [TimesformerModel](/docs/transformers/v5.23.0/ko/model_doc/timesformer#transformers.TimesformerModel) (TimesformerConfig model) - [TimmBackboneConfig](/docs/transformers/v5.23.0/ko/main_classes/backbones#transformers.TimmBackboneConfig) configuration class: [TimmBackbone](/docs/transformers/v5.23.0/ko/main_classes/backbones#transformers.TimmBackbone) (TimmBackboneConfig model) - [TimmWrapperConfig](/docs/transformers/v5.23.0/ko/model_doc/timm_wrapper#transformers.TimmWrapperConfig) configuration class: [TimmWrapperModel](/docs/transformers/v5.23.0/ko/model_doc/timm_wrapper#transformers.TimmWrapperModel) (TimmWrapperConfig model) - [Tipsv2Config](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2Config) configuration class: [Tipsv2Model](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2Model) (Tipsv2Config model) - [Tipsv2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2TextConfig) configuration class: [Tipsv2TextModel](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2TextModel) (Tipsv2TextConfig model) - [Tipsv2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2VisionConfig) configuration class: [Tipsv2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2VisionModel) (Tipsv2VisionConfig model) - [TvpConfig](/docs/transformers/v5.23.0/ko/model_doc/tvp#transformers.TvpConfig) configuration class: [TvpModel](/docs/transformers/v5.23.0/ko/model_doc/tvp#transformers.TvpModel) (TvpConfig model) - [UMT5Config](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5Config) configuration class: [UMT5Model](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5Model) (UMT5Config model) - [UVDocConfig](/docs/transformers/v5.23.0/ko/model_doc/uvdoc#transformers.UVDocConfig) configuration class: [UVDocModel](/docs/transformers/v5.23.0/ko/model_doc/uvdoc#transformers.UVDocModel) (UVDocConfig model) - [UdopConfig](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopConfig) configuration class: [UdopModel](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopModel) (UdopConfig model) - [UltraVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertConfig) configuration class: [UltraVBertModel](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertModel) (UltraVBertConfig model) - [UltrabertConfig](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertConfig) configuration class: [UltrabertModel](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertModel) (UltrabertConfig model) - [UniSpeechConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechConfig) configuration class: [UniSpeechModel](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechModel) (UniSpeechConfig model) - [UniSpeechSatConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatConfig) configuration class: [UniSpeechSatModel](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatModel) (UniSpeechSatConfig model) - [UnivNetConfig](/docs/transformers/v5.23.0/ko/model_doc/univnet#transformers.UnivNetConfig) configuration class: [UnivNetModel](/docs/transformers/v5.23.0/ko/model_doc/univnet#transformers.UnivNetModel) (UnivNetConfig model) - [UnlimitedOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrConfig) configuration class: [UnlimitedOcrModel](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrModel) (UnlimitedOcrConfig model) - [VJEPA2Config](/docs/transformers/v5.23.0/ko/model_doc/vjepa2#transformers.VJEPA2Config) configuration class: [VJEPA2Model](/docs/transformers/v5.23.0/ko/model_doc/vjepa2#transformers.VJEPA2Model) (VJEPA2Config model) - [VaultGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/vaultgemma#transformers.VaultGemmaConfig) configuration class: [VaultGemmaModel](/docs/transformers/v5.23.0/ko/model_doc/vaultgemma#transformers.VaultGemmaModel) (VaultGemmaConfig model) - [ViTConfig](/docs/transformers/v5.23.0/ko/model_doc/vit#transformers.ViTConfig) configuration class: [ViTModel](/docs/transformers/v5.23.0/ko/model_doc/vit#transformers.ViTModel) (ViTConfig model) - [ViTMAEConfig](/docs/transformers/v5.23.0/ko/model_doc/vit_mae#transformers.ViTMAEConfig) configuration class: [ViTMAEModel](/docs/transformers/v5.23.0/ko/model_doc/vit_mae#transformers.ViTMAEModel) (ViTMAEConfig model) - [ViTMSNConfig](/docs/transformers/v5.23.0/ko/model_doc/vit_msn#transformers.ViTMSNConfig) configuration class: [ViTMSNModel](/docs/transformers/v5.23.0/ko/model_doc/vit_msn#transformers.ViTMSNModel) (ViTMSNConfig model) - [VibeVoiceAcousticTokenizerConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerConfig) configuration class: [VibeVoiceAcousticTokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerModel) (VibeVoiceAcousticTokenizerConfig model) - [VibeVoiceAcousticTokenizerDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerDecoderConfig) configuration class: [VibeVoiceAcousticTokenizerDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerDecoderModel) (VibeVoiceAcousticTokenizerDecoderConfig model) - [VibeVoiceAcousticTokenizerEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerEncoderConfig) configuration class: [VibeVoiceAcousticTokenizerEncoderModel](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerEncoderModel) (VibeVoiceAcousticTokenizerEncoderConfig model) - [VibeVoiceAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrConfig) configuration class: [VibeVoiceAsrModel](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrModel) (VibeVoiceAsrConfig model) - [VibeVoiceConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice#transformers.VibeVoiceConfig) configuration class: [VibeVoiceForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice#transformers.VibeVoiceForConditionalGeneration) (VibeVoiceConfig model) - [VideoLlama3Config](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3Config) configuration class: [VideoLlama3Model](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3Model) (VideoLlama3Config model) - [VideoLlama3VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3VisionConfig) configuration class: [VideoLlama3VisionModel](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3VisionModel) (VideoLlama3VisionConfig model) - [VideoLlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaConfig) configuration class: [VideoLlavaModel](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaModel) (VideoLlavaConfig model) - [VideoMAEConfig](/docs/transformers/v5.23.0/ko/model_doc/videomae#transformers.VideoMAEConfig) configuration class: [VideoMAEModel](/docs/transformers/v5.23.0/ko/model_doc/videomae#transformers.VideoMAEModel) (VideoMAEConfig model) - [VideoPrismConfig](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismConfig) configuration class: [VideoPrismClipModel](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismClipModel) (VideoPrismConfig model) - [VideoPrismVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismVisionConfig) configuration class: [VideoPrismVisionModel](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismVisionModel) (VideoPrismVisionConfig model) - [ViltConfig](/docs/transformers/v5.23.0/ko/model_doc/vilt#transformers.ViltConfig) configuration class: [ViltModel](/docs/transformers/v5.23.0/ko/model_doc/vilt#transformers.ViltModel) (ViltConfig model) - [VipLlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaConfig) configuration class: [VipLlavaModel](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaModel) (VipLlavaConfig model) - [VisionTextDualEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/vision-text-dual-encoder#transformers.VisionTextDualEncoderConfig) configuration class: [VisionTextDualEncoderModel](/docs/transformers/v5.23.0/ko/model_doc/vision-text-dual-encoder#transformers.VisionTextDualEncoderModel) (VisionTextDualEncoderConfig model) - [VisualBertConfig](/docs/transformers/v5.23.0/ko/model_doc/visual_bert#transformers.VisualBertConfig) configuration class: [VisualBertModel](/docs/transformers/v5.23.0/ko/model_doc/visual_bert#transformers.VisualBertModel) (VisualBertConfig model) - [VitDetConfig](/docs/transformers/v5.23.0/ko/model_doc/vitdet#transformers.VitDetConfig) configuration class: [VitDetModel](/docs/transformers/v5.23.0/ko/model_doc/vitdet#transformers.VitDetModel) (VitDetConfig model) - [VitsConfig](/docs/transformers/v5.23.0/ko/model_doc/vits#transformers.VitsConfig) configuration class: [VitsModel](/docs/transformers/v5.23.0/ko/model_doc/vits#transformers.VitsModel) (VitsConfig model) - [VivitConfig](/docs/transformers/v5.23.0/ko/model_doc/vivit#transformers.VivitConfig) configuration class: [VivitModel](/docs/transformers/v5.23.0/ko/model_doc/vivit#transformers.VivitModel) (VivitConfig model) - [VoxtralConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralConfig) configuration class: [VoxtralModel](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralModel) (VoxtralConfig model) - [VoxtralEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralEncoderConfig) configuration class: [VoxtralEncoder](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralEncoder) (VoxtralEncoderConfig model) - [VoxtralRealtimeConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeConfig) configuration class: [VoxtralRealtimeModel](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeModel) (VoxtralRealtimeConfig model) - [VoxtralRealtimeEncoderConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeEncoderConfig) configuration class: [VoxtralRealtimeEncoder](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeEncoder) (VoxtralRealtimeEncoderConfig model) - [VoxtralRealtimeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeTextConfig) configuration class: `VoxtralRealtimeTextModel` (VoxtralRealtimeTextConfig model) - [Wav2Vec2BertConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertConfig) configuration class: [Wav2Vec2BertModel](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertModel) (Wav2Vec2BertConfig model) - [Wav2Vec2Config](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Config) configuration class: [Wav2Vec2Model](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Model) (Wav2Vec2Config model) - [Wav2Vec2ConformerConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerConfig) configuration class: [Wav2Vec2ConformerModel](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerModel) (Wav2Vec2ConformerConfig model) - [WavLMConfig](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMConfig) configuration class: [WavLMModel](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMModel) (WavLMConfig model) - [WhisperConfig](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperConfig) configuration class: [WhisperModel](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperModel) (WhisperConfig model) - [XCLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/xclip#transformers.XCLIPConfig) configuration class: [XCLIPModel](/docs/transformers/v5.23.0/ko/model_doc/xclip#transformers.XCLIPModel) (XCLIPConfig model) - [XGLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xglm#transformers.XGLMConfig) configuration class: [XGLMModel](/docs/transformers/v5.23.0/ko/model_doc/xglm#transformers.XGLMModel) (XGLMConfig model) - [XLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMConfig) configuration class: [XLMModel](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMModel) (XLMConfig model) - [XLMRobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaConfig) configuration class: [XLMRobertaModel](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaModel) (XLMRobertaConfig model) - [XLMRobertaXLConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLConfig) configuration class: [XLMRobertaXLModel](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLModel) (XLMRobertaXLConfig model) - [XLNetConfig](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetConfig) configuration class: [XLNetModel](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetModel) (XLNetConfig model) - [Xcodec2Config](/docs/transformers/v5.23.0/ko/model_doc/xcodec2#transformers.Xcodec2Config) configuration class: [Xcodec2Model](/docs/transformers/v5.23.0/ko/model_doc/xcodec2#transformers.Xcodec2Model) (Xcodec2Config model) - [XcodecConfig](/docs/transformers/v5.23.0/ko/model_doc/xcodec#transformers.XcodecConfig) configuration class: [XcodecModel](/docs/transformers/v5.23.0/ko/model_doc/xcodec#transformers.XcodecModel) (XcodecConfig model) - [XmodConfig](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodConfig) configuration class: [XmodModel](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodModel) (XmodConfig model) - [YolosConfig](/docs/transformers/v5.23.0/ko/model_doc/yolos#transformers.YolosConfig) configuration class: [YolosModel](/docs/transformers/v5.23.0/ko/model_doc/yolos#transformers.YolosModel) (YolosConfig model) - [YosoConfig](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoConfig) configuration class: [YosoModel](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoModel) (YosoConfig model) - [YoutuConfig](/docs/transformers/v5.23.0/ko/model_doc/youtu#transformers.YoutuConfig) configuration class: [YoutuModel](/docs/transformers/v5.23.0/ko/model_doc/youtu#transformers.YoutuModel) (YoutuConfig model) - [Zamba2Config](/docs/transformers/v5.23.0/ko/model_doc/zamba2#transformers.Zamba2Config) configuration class: [Zamba2Model](/docs/transformers/v5.23.0/ko/model_doc/zamba2#transformers.Zamba2Model) (Zamba2Config model) - [ZambaConfig](/docs/transformers/v5.23.0/ko/model_doc/zamba#transformers.ZambaConfig) configuration class: [ZambaModel](/docs/transformers/v5.23.0/ko/model_doc/zamba#transformers.ZambaModel) (ZambaConfig model) - [ZayaConfig](/docs/transformers/v5.23.0/ko/model_doc/zaya#transformers.ZayaConfig) configuration class: [ZayaModel](/docs/transformers/v5.23.0/ko/model_doc/zaya#transformers.ZayaModel) (ZayaConfig model) - [xLSTMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlstm#transformers.xLSTMConfig) configuration class: [xLSTMModel](/docs/transformers/v5.23.0/ko/model_doc/xlstm#transformers.xLSTMModel) (xLSTMConfig model)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

configuration에서 라이브러리의 기본 모델 클래스 중 하나를 초기화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModel

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModel.from_config(config)
```

#### from_pretrained[[transformers.AutoModel.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **afmoe** -- [AfmoeModel](/docs/transformers/v5.23.0/ko/model_doc/afmoe#transformers.AfmoeModel) (AfmoeConfig 모델)
- **aimv2** -- [Aimv2Model](/docs/transformers/v5.23.0/ko/model_doc/aimv2#transformers.Aimv2Model) (Aimv2Config 모델)
- **aimv2_vision_model** -- [Aimv2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/aimv2#transformers.Aimv2VisionModel) (Aimv2VisionConfig 모델)
- **albert** -- [AlbertModel](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertModel) (AlbertConfig 모델)
- **align** -- [AlignModel](/docs/transformers/v5.23.0/ko/model_doc/align#transformers.AlignModel) (AlignConfig 모델)
- **altclip** -- [AltCLIPModel](/docs/transformers/v5.23.0/ko/model_doc/altclip#transformers.AltCLIPModel) (AltCLIPConfig 모델)
- **apertus** -- [ApertusModel](/docs/transformers/v5.23.0/ko/model_doc/apertus#transformers.ApertusModel) (ApertusConfig 모델)
- **arcee** -- [ArceeModel](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeModel) (ArceeConfig 모델)
- **aria** -- [AriaModel](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaModel) (AriaConfig 모델)
- **aria_text** -- [AriaTextModel](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaTextModel) (AriaTextConfig 모델)
- **audio-spectrogram-transformer** -- [ASTModel](/docs/transformers/v5.23.0/ko/model_doc/audio-spectrogram-transformer#transformers.ASTModel) (ASTConfig 모델)
- **audioflamingo3** -- [AudioFlamingo3Model](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3Model) (AudioFlamingo3Config 모델)
- **audioflamingo3_encoder** -- [AudioFlamingo3Encoder](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3Encoder) (AudioFlamingo3EncoderConfig 모델)
- **autoformer** -- [AutoformerModel](/docs/transformers/v5.23.0/ko/model_doc/autoformer#transformers.AutoformerModel) (AutoformerConfig 모델)
- **axk1** -- [AXK1Model](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1Model) (AXK1Config 모델)
- **axk2** -- [AXK2Model](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2Model) (AXK2Config 모델)
- **aya_vision** -- [AyaVisionModel](/docs/transformers/v5.23.0/ko/model_doc/aya_vision#transformers.AyaVisionModel) (AyaVisionConfig 모델)
- **bailing_hybrid** -- [BailingHybridModel](/docs/transformers/v5.23.0/ko/model_doc/bailing_hybrid#transformers.BailingHybridModel) (BailingHybridConfig 모델)
- **bamba** -- [BambaModel](/docs/transformers/v5.23.0/ko/model_doc/bamba#transformers.BambaModel) (BambaConfig 모델)
- **bark** -- [BarkModel](/docs/transformers/v5.23.0/ko/model_doc/bark#transformers.BarkModel) (BarkConfig 모델)
- **bart** -- [BartModel](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartModel) (BartConfig 모델)
- **beam** -- [BeamModel](/docs/transformers/v5.23.0/ko/model_doc/beam#transformers.BeamModel) (BeamConfig 모델)
- **beit** -- [BeitModel](/docs/transformers/v5.23.0/ko/model_doc/beit#transformers.BeitModel) (BeitConfig 모델)
- **bert** -- [BertModel](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertModel) (BertConfig 모델)
- **bert-generation** -- [BertGenerationEncoder](/docs/transformers/v5.23.0/ko/model_doc/bert-generation#transformers.BertGenerationEncoder) (BertGenerationConfig 모델)
- **big_bird** -- [BigBirdModel](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdModel) (BigBirdConfig 모델)
- **bigbird_pegasus** -- [BigBirdPegasusModel](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusModel) (BigBirdPegasusConfig 모델)
- **biogpt** -- [BioGptModel](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptModel) (BioGptConfig 모델)
- **bit** -- [BitModel](/docs/transformers/v5.23.0/ko/model_doc/bit#transformers.BitModel) (BitConfig 모델)
- **bitnet** -- [BitNetModel](/docs/transformers/v5.23.0/ko/model_doc/bitnet#transformers.BitNetModel) (BitNetConfig 모델)
- **blenderbot** -- [BlenderbotModel](/docs/transformers/v5.23.0/ko/model_doc/blenderbot#transformers.BlenderbotModel) (BlenderbotConfig 모델)
- **blenderbot-small** -- [BlenderbotSmallModel](/docs/transformers/v5.23.0/ko/model_doc/blenderbot-small#transformers.BlenderbotSmallModel) (BlenderbotSmallConfig 모델)
- **blip** -- [BlipModel](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipModel) (BlipConfig 모델)
- **blip-2** -- [Blip2Model](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2Model) (Blip2Config 모델)
- **blip_2_qformer** -- [Blip2QFormerModel](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2QFormerModel) (Blip2QFormerConfig 모델)
- **bloom** -- [BloomModel](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomModel) (BloomConfig 모델)
- **blt** -- [BltModel](/docs/transformers/v5.23.0/ko/model_doc/blt#transformers.BltModel) (BltConfig 모델)
- **bridgetower** -- [BridgeTowerModel](/docs/transformers/v5.23.0/ko/model_doc/bridgetower#transformers.BridgeTowerModel) (BridgeTowerConfig 모델)
- **bros** -- [BrosModel](/docs/transformers/v5.23.0/ko/model_doc/bros#transformers.BrosModel) (BrosConfig 모델)
- **camembert** -- [CamembertModel](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertModel) (CamembertConfig 모델)
- **canary** -- [CanaryModel](/docs/transformers/v5.23.0/ko/model_doc/canary#transformers.CanaryModel) (CanaryConfig 모델)
- **canine** -- [CanineModel](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineModel) (CanineConfig 모델)
- **chameleon** -- [ChameleonModel](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonModel) (ChameleonConfig 모델)
- **chinese_clip** -- [ChineseCLIPModel](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPModel) (ChineseCLIPConfig 모델)
- **chinese_clip_vision_model** -- [ChineseCLIPVisionModel](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPVisionModel) (ChineseCLIPVisionConfig 모델)
- **clap** -- [ClapModel](/docs/transformers/v5.23.0/ko/model_doc/clap#transformers.ClapModel) (ClapConfig 모델)
- **clip** -- [CLIPModel](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPModel) (CLIPConfig 모델)
- **clip_text_model** -- [CLIPTextModel](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPTextModel) (CLIPTextConfig 모델)
- **clip_vision_model** -- [CLIPVisionModel](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPVisionModel) (CLIPVisionConfig 모델)
- **clipseg** -- [CLIPSegModel](/docs/transformers/v5.23.0/ko/model_doc/clipseg#transformers.CLIPSegModel) (CLIPSegConfig 모델)
- **clvp** -- [ClvpModelForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/clvp#transformers.ClvpModelForConditionalGeneration) (ClvpConfig 모델)
- **codegen** -- [CodeGenModel](/docs/transformers/v5.23.0/ko/model_doc/codegen#transformers.CodeGenModel) (CodeGenConfig 모델)
- **cohere** -- [CohereModel](/docs/transformers/v5.23.0/ko/model_doc/cohere#transformers.CohereModel) (CohereConfig 모델)
- **cohere2** -- [Cohere2Model](/docs/transformers/v5.23.0/ko/model_doc/cohere2#transformers.Cohere2Model) (Cohere2Config 모델)
- **cohere2_moe** -- [Cohere2MoeModel](/docs/transformers/v5.23.0/ko/model_doc/cohere2_moe#transformers.Cohere2MoeModel) (Cohere2MoeConfig 모델)
- **cohere2_vision** -- [Cohere2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/cohere2_vision#transformers.Cohere2VisionModel) (Cohere2VisionConfig 모델)
- **cohere_asr** -- [CohereAsrModel](/docs/transformers/v5.23.0/ko/model_doc/cohere_asr#transformers.CohereAsrModel) (CohereAsrConfig 모델)
- **cohere_compass** -- [CohereCompassModel](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassModel) (CohereCompassConfig 모델)
- **cohere_compass_text** -- [CohereCompassTextModel](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassTextModel) (CohereCompassTextConfig 모델)
- **cohere_compass_vision** -- [CohereCompassVisionModel](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassVisionModel) (CohereCompassVisionConfig 모델)
- **conditional_detr** -- [ConditionalDetrModel](/docs/transformers/v5.23.0/ko/model_doc/conditional_detr#transformers.ConditionalDetrModel) (ConditionalDetrConfig 모델)
- **convbert** -- [ConvBertModel](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertModel) (ConvBertConfig 모델)
- **convnext** -- [ConvNextModel](/docs/transformers/v5.23.0/ko/model_doc/convnext#transformers.ConvNextModel) (ConvNextConfig 모델)
- **convnextv2** -- [ConvNextV2Model](/docs/transformers/v5.23.0/ko/model_doc/convnextv2#transformers.ConvNextV2Model) (ConvNextV2Config 모델)
- **cosmos3_edge** -- [Cosmos3EdgeModel](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeModel) (Cosmos3EdgeConfig 모델)
- **cosmos3_edge_text** -- [Cosmos3EdgeTextModel](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeTextModel) (Cosmos3EdgeTextConfig 모델)
- **cosmos3_edge_vision** -- [Cosmos3EdgeVisionModel](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeVisionModel) (Cosmos3EdgeVisionConfig 모델)
- **cosmos3_omni** -- [Cosmos3OmniModel](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_omni#transformers.Cosmos3OmniModel) (Cosmos3OmniConfig 모델)
- **cpmant** -- [CpmAntModel](/docs/transformers/v5.23.0/ko/model_doc/cpmant#transformers.CpmAntModel) (CpmAntConfig 모델)
- **csm** -- [CsmForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/csm#transformers.CsmForConditionalGeneration) (CsmConfig 모델)
- **ctrl** -- [CTRLModel](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLModel) (CTRLConfig 모델)
- **cvt** -- [CvtModel](/docs/transformers/v5.23.0/ko/model_doc/cvt#transformers.CvtModel) (CvtConfig 모델)
- **cwm** -- [CwmModel](/docs/transformers/v5.23.0/ko/model_doc/cwm#transformers.CwmModel) (CwmConfig 모델)
- **d_fine** -- [DFineModel](/docs/transformers/v5.23.0/ko/model_doc/d_fine#transformers.DFineModel) (DFineConfig 모델)
- **dab-detr** -- [DabDetrModel](/docs/transformers/v5.23.0/ko/model_doc/dab-detr#transformers.DabDetrModel) (DabDetrConfig 모델)
- **dac** -- [DacModel](/docs/transformers/v5.23.0/ko/model_doc/dac#transformers.DacModel) (DacConfig 모델)
- **data2vec-audio** -- [Data2VecAudioModel](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioModel) (Data2VecAudioConfig 모델)
- **data2vec-text** -- [Data2VecTextModel](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextModel) (Data2VecTextConfig 모델)
- **data2vec-vision** -- [Data2VecVisionModel](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecVisionModel) (Data2VecVisionConfig 모델)
- **dbrx** -- [DbrxModel](/docs/transformers/v5.23.0/ko/model_doc/dbrx#transformers.DbrxModel) (DbrxConfig 모델)
- **deberta** -- [DebertaModel](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaModel) (DebertaConfig 모델)
- **deberta-v2** -- [DebertaV2Model](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2Model) (DebertaV2Config 모델)
- **decision_transformer** -- [DecisionTransformerModel](/docs/transformers/v5.23.0/ko/model_doc/decision_transformer#transformers.DecisionTransformerModel) (DecisionTransformerConfig 모델)
- **deepseek_ocr2** -- [DeepseekOcr2Model](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2Model) (DeepseekOcr2Config 모델)
- **deepseek_v2** -- [DeepseekV2Model](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v2#transformers.DeepseekV2Model) (DeepseekV2Config 모델)
- **deepseek_v3** -- [DeepseekV3Model](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3Model) (DeepseekV3Config 모델)
- **deepseek_v32** -- [DeepseekV32Model](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v32#transformers.DeepseekV32Model) (DeepseekV32Config 모델)
- **deepseek_v4** -- [DeepseekV4Model](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v4#transformers.DeepseekV4Model) (DeepseekV4Config 모델)
- **deepseek_vl** -- [DeepseekVLModel](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl#transformers.DeepseekVLModel) (DeepseekVLConfig 모델)
- **deepseek_vl_hybrid** -- [DeepseekVLHybridModel](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl_hybrid#transformers.DeepseekVLHybridModel) (DeepseekVLHybridConfig 모델)
- **deformable_detr** -- [DeformableDetrModel](/docs/transformers/v5.23.0/ko/model_doc/deformable_detr#transformers.DeformableDetrModel) (DeformableDetrConfig 모델)
- **deimv2** -- [Deimv2Model](/docs/transformers/v5.23.0/ko/model_doc/deimv2#transformers.Deimv2Model) (Deimv2Config 모델)
- **deit** -- [DeiTModel](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTModel) (DeiTConfig 모델)
- **depth_pro** -- [DepthProModel](/docs/transformers/v5.23.0/ko/model_doc/depth_pro#transformers.DepthProModel) (DepthProConfig 모델)
- **detr** -- [DetrModel](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrModel) (DetrConfig 모델)
- **dia** -- [DiaModel](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaModel) (DiaConfig 모델)
- **diffllama** -- [DiffLlamaModel](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaModel) (DiffLlamaConfig 모델)
- **diffusion_gemma** -- [DiffusionGemmaModel](/docs/transformers/v5.23.0/ko/model_doc/diffusion_gemma#transformers.DiffusionGemmaModel) (DiffusionGemmaConfig 모델)
- **dinat** -- [DinatModel](/docs/transformers/v5.23.0/ko/model_doc/dinat#transformers.DinatModel) (DinatConfig 모델)
- **dinov2** -- [Dinov2Model](/docs/transformers/v5.23.0/ko/model_doc/dinov2#transformers.Dinov2Model) (Dinov2Config 모델)
- **dinov2_with_registers** -- [Dinov2WithRegistersModel](/docs/transformers/v5.23.0/ko/model_doc/dinov2_with_registers#transformers.Dinov2WithRegistersModel) (Dinov2WithRegistersConfig 모델)
- **dinov3_convnext** -- [DINOv3ConvNextModel](/docs/transformers/v5.23.0/ko/model_doc/dinov3#transformers.DINOv3ConvNextModel) (DINOv3ConvNextConfig 모델)
- **dinov3_vit** -- [DINOv3ViTModel](/docs/transformers/v5.23.0/ko/model_doc/dinov3#transformers.DINOv3ViTModel) (DINOv3ViTConfig 모델)
- **distilbert** -- [DistilBertModel](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertModel) (DistilBertConfig 모델)
- **doge** -- [DogeModel](/docs/transformers/v5.23.0/ko/model_doc/doge#transformers.DogeModel) (DogeConfig 모델)
- **donut-swin** -- [DonutSwinModel](/docs/transformers/v5.23.0/ko/model_doc/donut#transformers.DonutSwinModel) (DonutSwinConfig 모델)
- **dots1** -- [Dots1Model](/docs/transformers/v5.23.0/ko/model_doc/dots1#transformers.Dots1Model) (Dots1Config 모델)
- **dots3_note** -- [Dots3NoteModel](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteModel) (Dots3NoteConfig 모델)
- **dots3_note_text** -- [Dots3NoteTextModel](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteTextModel) (Dots3NoteTextConfig 모델)
- **dpr** -- [DPRQuestionEncoder](/docs/transformers/v5.23.0/ko/model_doc/dpr#transformers.DPRQuestionEncoder) (DPRConfig 모델)
- **dpt** -- [DPTModel](/docs/transformers/v5.23.0/ko/model_doc/dpt#transformers.DPTModel) (DPTConfig 모델)
- **edgetam** -- [EdgeTamModel](/docs/transformers/v5.23.0/ko/model_doc/edgetam#transformers.EdgeTamModel) (EdgeTamConfig 모델)
- **edgetam_video** -- [EdgeTamVideoModel](/docs/transformers/v5.23.0/ko/model_doc/edgetam_video#transformers.EdgeTamVideoModel) (EdgeTamVideoConfig 모델)
- **edgetam_vision_model** -- [EdgeTamVisionModel](/docs/transformers/v5.23.0/ko/model_doc/edgetam#transformers.EdgeTamVisionModel) (EdgeTamVisionConfig 모델)
- **efficientloftr** -- [EfficientLoFTRModel](/docs/transformers/v5.23.0/ko/model_doc/efficientloftr#transformers.EfficientLoFTRModel) (EfficientLoFTRConfig 모델)
- **efficientnet** -- [EfficientNetModel](/docs/transformers/v5.23.0/ko/model_doc/efficientnet#transformers.EfficientNetModel) (EfficientNetConfig 모델)
- **electra** -- [ElectraModel](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraModel) (ElectraConfig 모델)
- **embedding_gemma2** -- [EmbeddingGemma2Model](/docs/transformers/v5.23.0/ko/model_doc/embedding_gemma2#transformers.EmbeddingGemma2Model) (EmbeddingGemma2Config 모델)
- **embedding_gemma2_text** -- [EmbeddingGemma2TextModel](/docs/transformers/v5.23.0/ko/model_doc/embedding_gemma2#transformers.EmbeddingGemma2TextModel) (EmbeddingGemma2TextConfig 모델)
- **emu3** -- [Emu3Model](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3Model) (Emu3Config 모델)
- **encodec** -- [EncodecModel](/docs/transformers/v5.23.0/ko/model_doc/encodec#transformers.EncodecModel) (EncodecConfig 모델)
- **ernie** -- [ErnieModel](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieModel) (ErnieConfig 모델)
- **ernie4_5** -- [Ernie4_5Model](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5#transformers.Ernie4_5Model) (Ernie4_5Config 모델)
- **ernie4_5_moe** -- [Ernie4_5_MoeModel](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_moe#transformers.Ernie4_5_MoeModel) (Ernie4_5_MoeConfig 모델)
- **ernie4_5_vl_moe** -- [Ernie4_5_VLMoeModel](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeModel) (Ernie4_5_VLMoeConfig 모델)
- **esm** -- [EsmModel](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmModel) (EsmConfig 모델)
- **esmc** -- [EsmcModel](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcModel) (EsmcConfig 모델)
- **eurobert** -- [EuroBertModel](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertModel) (EuroBertConfig 모델)
- **evolla** -- [EvollaModel](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaModel) (EvollaConfig 모델)
- **exaone4** -- [Exaone4Model](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4Model) (Exaone4Config 모델)
- **exaone4_5** -- [Exaone4_5_Model](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_Model) (Exaone4_5_Config 모델)
- **exaone4_5_vision** -- [Exaone4_5_VisionModel](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_VisionModel) (Exaone4_5_VisionConfig 모델)
- **exaone_moe** -- [ExaoneMoeModel](/docs/transformers/v5.23.0/ko/model_doc/exaone_moe#transformers.ExaoneMoeModel) (ExaoneMoeConfig 모델)
- **falcon** -- [FalconModel](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconModel) (FalconConfig 모델)
- **falcon_h1** -- [FalconH1Model](/docs/transformers/v5.23.0/ko/model_doc/falcon_h1#transformers.FalconH1Model) (FalconH1Config 모델)
- **falcon_mamba** -- [FalconMambaModel](/docs/transformers/v5.23.0/ko/model_doc/falcon_mamba#transformers.FalconMambaModel) (FalconMambaConfig 모델)
- **fast_vlm** -- [FastVlmModel](/docs/transformers/v5.23.0/ko/model_doc/fast_vlm#transformers.FastVlmModel) (FastVlmConfig 모델)
- **fastspeech2_conformer** -- [FastSpeech2ConformerModel](/docs/transformers/v5.23.0/ko/model_doc/fastspeech2_conformer#transformers.FastSpeech2ConformerModel) (FastSpeech2ConformerConfig 모델)
- **fastspeech2_conformer_with_hifigan** -- [FastSpeech2ConformerWithHifiGan](/docs/transformers/v5.23.0/ko/model_doc/fastspeech2_conformer#transformers.FastSpeech2ConformerWithHifiGan) (FastSpeech2ConformerWithHifiGanConfig 모델)
- **flaubert** -- [FlaubertModel](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertModel) (FlaubertConfig 모델)
- **flava** -- [FlavaModel](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaModel) (FlavaConfig 모델)
- **flex_olmo** -- [FlexOlmoModel](/docs/transformers/v5.23.0/ko/model_doc/flex_olmo#transformers.FlexOlmoModel) (FlexOlmoConfig 모델)
- **flint** -- [FlintModel](/docs/transformers/v5.23.0/ko/model_doc/flint#transformers.FlintModel) (FlintConfig 모델)
- **florence2** -- [Florence2Model](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2Model) (Florence2Config 모델)
- **fnet** -- [FNetModel](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetModel) (FNetConfig 모델)
- **focalnet** -- [FocalNetModel](/docs/transformers/v5.23.0/ko/model_doc/focalnet#transformers.FocalNetModel) (FocalNetConfig 모델)
- **fsmt** -- [FSMTModel](/docs/transformers/v5.23.0/ko/model_doc/fsmt#transformers.FSMTModel) (FSMTConfig 모델)
- **fun_asr_nano** -- [FunAsrNanoModel](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoModel) (FunAsrNanoConfig 모델)
- **fun_asr_nano_encoder** -- [FunAsrNanoEncoder](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoEncoder) (FunAsrNanoEncoderConfig 모델)
- **funnel** -- [FunnelModel](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelModel) 또는 [FunnelBaseModel](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelBaseModel) (FunnelConfig 모델)
- **fuyu** -- [FuyuModel](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuModel) (FuyuConfig 모델)
- **gemma** -- [GemmaModel](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaModel) (GemmaConfig 모델)
- **gemma2** -- [Gemma2Model](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2Model) (Gemma2Config 모델)
- **gemma3** -- [Gemma3Model](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Model) (Gemma3Config 모델)
- **gemma3_text** -- [Gemma3TextModel](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3TextModel) (Gemma3TextConfig 모델)
- **gemma3n** -- [Gemma3nModel](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nModel) (Gemma3nConfig 모델)
- **gemma3n_audio** -- `Gemma3nAudioEncoder` (Gemma3nAudioConfig 모델)
- **gemma3n_text** -- [Gemma3nTextModel](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nTextModel) (Gemma3nTextConfig 모델)
- **gemma3n_vision** -- [TimmWrapperModel](/docs/transformers/v5.23.0/ko/model_doc/timm_wrapper#transformers.TimmWrapperModel) (Gemma3nVisionConfig 모델)
- **gemma4** -- [Gemma4Model](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4Model) (Gemma4Config 모델)
- **gemma4_audio** -- [Gemma4AudioModel](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4AudioModel) (Gemma4AudioConfig 모델)
- **gemma4_text** -- [Gemma4TextModel](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4TextModel) (Gemma4TextConfig 모델)
- **gemma4_unified** -- [Gemma4UnifiedModel](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedModel) (Gemma4UnifiedConfig 모델)
- **gemma4_unified_text** -- [Gemma4UnifiedTextModel](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedTextModel) (Gemma4UnifiedTextConfig 모델)
- **gemma4_vision** -- [Gemma4VisionModel](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4VisionModel) (Gemma4VisionConfig 모델)
- **git** -- [GitModel](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitModel) (GitConfig 모델)
- **glm** -- [GlmModel](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmModel) (GlmConfig 모델)
- **glm4** -- [Glm4Model](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4Model) (Glm4Config 모델)
- **glm46v** -- [Glm46VModel](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VModel) (Glm46VConfig 모델)
- **glm4_moe** -- [Glm4MoeModel](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe#transformers.Glm4MoeModel) (Glm4MoeConfig 모델)
- **glm4_moe_lite** -- [Glm4MoeLiteModel](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe_lite#transformers.Glm4MoeLiteModel) (Glm4MoeLiteConfig 모델)
- **glm4v** -- [Glm4vModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vModel) (Glm4vConfig 모델)
- **glm4v_moe** -- [Glm4vMoeModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeModel) (Glm4vMoeConfig 모델)
- **glm4v_moe_text** -- [Glm4vMoeTextModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeTextModel) (Glm4vMoeTextConfig 모델)
- **glm4v_moe_vision** -- [Glm4vMoeVisionModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeVisionModel) (Glm4vMoeVisionConfig 모델)
- **glm4v_text** -- [Glm4vTextModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vTextModel) (Glm4vTextConfig 모델)
- **glm4v_vision** -- [Glm4vVisionModel](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vVisionModel) (Glm4vVisionConfig 모델)
- **glm5_next** -- [Glm5NextModel](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextModel) (Glm5NextConfig 모델)
- **glm5_next_text** -- [Glm5NextTextModel](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextTextModel) (Glm5NextTextConfig 모델)
- **glm5_next_vision** -- [Glm5NextVisionModel](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextVisionModel) (Glm5NextVisionConfig 모델)
- **glm_image** -- [GlmImageModel](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageModel) (GlmImageConfig 모델)
- **glm_image_text** -- [GlmImageTextModel](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageTextModel) (GlmImageTextConfig 모델)
- **glm_image_vision** -- [GlmImageVisionModel](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageVisionModel) (GlmImageVisionConfig 모델)
- **glm_image_vqmodel** -- [GlmImageVQVAE](/docs/transformers/v5.23.0/ko/model_doc/glm_image#transformers.GlmImageVQVAE) (GlmImageVQVAEConfig 모델)
- **glm_moe_dsa** -- [GlmMoeDsaModel](/docs/transformers/v5.23.0/ko/model_doc/glm_moe_dsa#transformers.GlmMoeDsaModel) (GlmMoeDsaConfig 모델)
- **glm_ocr** -- [GlmOcrModel](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrModel) (GlmOcrConfig 모델)
- **glm_ocr_text** -- [GlmOcrTextModel](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrTextModel) (GlmOcrTextConfig 모델)
- **glm_ocr_vision** -- [GlmOcrVisionModel](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrVisionModel) (GlmOcrVisionConfig 모델)
- **glmasr** -- [GlmAsrModel](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrModel) (GlmAsrConfig 모델)
- **glmasr_encoder** -- [GlmAsrEncoder](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrEncoder) (GlmAsrEncoderConfig 모델)
- **glmga** -- [Glm46VModel](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VModel) (GlmgaConfig 모델)
- **glpn** -- [GLPNModel](/docs/transformers/v5.23.0/ko/model_doc/glpn#transformers.GLPNModel) (GLPNConfig 모델)
- **got_ocr2** -- [GotOcr2Model](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2Model) (GotOcr2Config 모델)
- **gpt-sw3** -- [GPT2Model](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Model) (GPT2Config 모델)
- **gpt2** -- [GPT2Model](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Model) (GPT2Config 모델)
- **gpt_bigcode** -- [GPTBigCodeModel](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeModel) (GPTBigCodeConfig 모델)
- **gpt_neo** -- [GPTNeoModel](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoModel) (GPTNeoConfig 모델)
- **gpt_neox** -- [GPTNeoXModel](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXModel) (GPTNeoXConfig 모델)
- **gpt_neox_japanese** -- [GPTNeoXJapaneseModel](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox_japanese#transformers.GPTNeoXJapaneseModel) (GPTNeoXJapaneseConfig 모델)
- **gpt_oss** -- [GptOssModel](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssModel) (GptOssConfig 모델)
- **gptj** -- [GPTJModel](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJModel) (GPTJConfig 모델)
- **gradium_phonon_tts** -- [GradiumPhononTTSModel](/docs/transformers/v5.23.0/ko/model_doc/gradium_phonon_tts#transformers.GradiumPhononTTSModel) (GradiumPhononTTSConfig 모델)
- **granite** -- [GraniteModel](/docs/transformers/v5.23.0/ko/model_doc/granite#transformers.GraniteModel) (GraniteConfig 모델)
- **granite4_vision** -- [Granite4VisionModel](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionModel) (Granite4VisionConfig 모델)
- **granite_docling** -- [GraniteDoclingModel](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingModel) (GraniteDoclingConfig 모델)
- **granite_docling_text** -- [GraniteDoclingTextModel](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingTextModel) (GraniteDoclingTextConfig 모델)
- **granite_docling_vision** -- [GraniteDoclingVisionModel](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingVisionModel) (GraniteDoclingVisionConfig 모델)
- **granite_speech** -- [GraniteSpeechModel](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechModel) (GraniteSpeechConfig 모델)
- **granite_speech5_ctc** -- [GraniteSpeech5ForCTC](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5ForCTC) (GraniteSpeech5CTCConfig 모델)
- **granite_speech5_encoder** -- [GraniteSpeech5Encoder](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5Encoder) (GraniteSpeech5EncoderConfig 모델)
- **granite_speech_plus** -- [GraniteSpeechPlusModel](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusModel) (GraniteSpeechPlusConfig 모델)
- **granite_swa** -- [GraniteSWAModel](/docs/transformers/v5.23.0/ko/model_doc/granite_swa#transformers.GraniteSWAModel) (GraniteSWAConfig 모델)
- **granitemoe** -- [GraniteMoeModel](/docs/transformers/v5.23.0/ko/model_doc/granitemoe#transformers.GraniteMoeModel) (GraniteMoeConfig 모델)
- **granitemoe_swa** -- [GraniteMoeSWAModel](/docs/transformers/v5.23.0/ko/model_doc/granitemoe_swa#transformers.GraniteMoeSWAModel) (GraniteMoeSWAConfig 모델)
- **granitemoehybrid** -- [GraniteMoeHybridModel](/docs/transformers/v5.23.0/ko/model_doc/granitemoehybrid#transformers.GraniteMoeHybridModel) (GraniteMoeHybridConfig 모델)
- **granitemoeshared** -- [GraniteMoeSharedModel](/docs/transformers/v5.23.0/ko/model_doc/granitemoeshared#transformers.GraniteMoeSharedModel) (GraniteMoeSharedConfig 모델)
- **grounding-dino** -- [GroundingDinoModel](/docs/transformers/v5.23.0/ko/model_doc/grounding-dino#transformers.GroundingDinoModel) (GroundingDinoConfig 모델)
- **groupvit** -- [GroupViTModel](/docs/transformers/v5.23.0/ko/model_doc/groupvit#transformers.GroupViTModel) (GroupViTConfig 모델)
- **gte** -- [GteModel](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteModel) (GteConfig 모델)
- **helium** -- [HeliumModel](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumModel) (HeliumConfig 모델)
- **hgnet_v2** -- [HGNetV2Backbone](/docs/transformers/v5.23.0/ko/model_doc/hgnet_v2#transformers.HGNetV2Backbone) (HGNetV2Config 모델)
- **hiera** -- [HieraModel](/docs/transformers/v5.23.0/ko/model_doc/hiera#transformers.HieraModel) (HieraConfig 모델)
- **higgs_audio_v2** -- [HiggsAudioV2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2#transformers.HiggsAudioV2ForConditionalGeneration) (HiggsAudioV2Config 모델)
- **higgs_audio_v2_tokenizer** -- [HiggsAudioV2TokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2_tokenizer#transformers.HiggsAudioV2TokenizerModel) (HiggsAudioV2TokenizerConfig 모델)
- **hrm_text** -- [HrmTextModel](/docs/transformers/v5.23.0/ko/model_doc/hrm_text#transformers.HrmTextModel) (HrmTextConfig 모델)
- **hubert** -- [HubertModel](/docs/transformers/v5.23.0/ko/model_doc/hubert#transformers.HubertModel) (HubertConfig 모델)
- **hunyuan_v1_dense** -- [HunYuanDenseV1Model](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_dense#transformers.HunYuanDenseV1Model) (HunYuanDenseV1Config 모델)
- **hunyuan_v1_moe** -- [HunYuanMoEV1Model](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_moe#transformers.HunYuanMoEV1Model) (HunYuanMoEV1Config 모델)
- **hunyuan_vl** -- [HunYuanVLModel](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLModel) (HunYuanVLConfig 모델)
- **hunyuan_vl_text** -- [HunYuanVLTextModel](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLTextModel) (HunYuanVLTextConfig 모델)
- **hunyuan_vl_vision** -- [HunYuanVLVisionTransformer](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLVisionTransformer) (HunYuanVLVisionConfig 모델)
- **hy_v3** -- [HYV3Model](/docs/transformers/v5.23.0/ko/model_doc/hy_v3#transformers.HYV3Model) (HYV3Config 모델)
- **hy_v4** -- [HYV4Model](/docs/transformers/v5.23.0/ko/model_doc/hy_v4#transformers.HYV4Model) (HYV4Config 모델)
- **hyperclovax** -- [HyperCLOVAXModel](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax#transformers.HyperCLOVAXModel) (HyperCLOVAXConfig 모델)
- **hyperclovax_vision_v2** -- [HyperCLOVAXVisionV2Model](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax_vision_v2#transformers.HyperCLOVAXVisionV2Model) (HyperCLOVAXVisionV2Config 모델)
- **ibert** -- [IBertModel](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertModel) (IBertConfig 모델)
- **idefics** -- [IdeficsModel](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsModel) (IdeficsConfig 모델)
- **idefics2** -- [Idefics2Model](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2Model) (Idefics2Config 모델)
- **idefics3** -- [Idefics3Model](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3Model) (Idefics3Config 모델)
- **idefics3_vision** -- [Idefics3VisionTransformer](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3VisionTransformer) (Idefics3VisionConfig 모델)
- **ijepa** -- [IJepaModel](/docs/transformers/v5.23.0/ko/model_doc/ijepa#transformers.IJepaModel) (IJepaConfig 모델)
- **imagegpt** -- [ImageGPTModel](/docs/transformers/v5.23.0/ko/model_doc/imagegpt#transformers.ImageGPTModel) (ImageGPTConfig 모델)
- **informer** -- [InformerModel](/docs/transformers/v5.23.0/ko/model_doc/informer#transformers.InformerModel) (InformerConfig 모델)
- **inkling2** -- [Inkling2Model](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2Model) (Inkling2Config 모델)
- **inkling2_text** -- [Inkling2TextModel](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2TextModel) (Inkling2TextConfig 모델)
- **inkling2_vision** -- [Inkling2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2VisionModel) (Inkling2VisionConfig 모델)
- **inkling_mm_model** -- [InklingModel](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingModel) (InklingConfig 모델)
- **inkling_text** -- [InklingTextModel](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingTextModel) (InklingTextConfig 모델)
- **instructblip** -- [InstructBlipModel](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipModel) (InstructBlipConfig 모델)
- **instructblipvideo** -- [InstructBlipVideoModel](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoModel) (InstructBlipVideoConfig 모델)
- **internlm3** -- [LlamaModel](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaModel) (InternLM3Config 모델)
- **internvl** -- [InternVLModel](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLModel) (InternVLConfig 모델)
- **internvl_vision** -- [InternVLVisionModel](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLVisionModel) (InternVLVisionConfig 모델)
- **iquestcoder** -- [LlamaModel](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaModel) (IQuestCoderConfig 모델)
- **jais2** -- [Jais2Model](/docs/transformers/v5.23.0/ko/model_doc/jais2#transformers.Jais2Model) (Jais2Config 모델)
- **jamba** -- [JambaModel](/docs/transformers/v5.23.0/ko/model_doc/jamba#transformers.JambaModel) (JambaConfig 모델)
- **janus** -- [JanusModel](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusModel) (JanusConfig 모델)
- **jetmoe** -- [JetMoeModel](/docs/transformers/v5.23.0/ko/model_doc/jetmoe#transformers.JetMoeModel) (JetMoeConfig 모델)
- **jina_embeddings_v3** -- [JinaEmbeddingsV3Model](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3Model) (JinaEmbeddingsV3Config 모델)
- **kimi_k25** -- [Kimi_K25Model](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25Model) (Kimi_K25Config 모델)
- **kimi_k25_vision** -- [Kimi_K25VisionModel](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25VisionModel) (Kimi_K25VisionConfig 모델)
- **kimi_linear** -- [KimiLinearModel](/docs/transformers/v5.23.0/ko/model_doc/kimi_linear#transformers.KimiLinearModel) (KimiLinearConfig 모델)
- **kolibri1** -- [Kolibri1Model](/docs/transformers/v5.23.0/ko/model_doc/kolibri1#transformers.Kolibri1Model) (Kolibri1Config 모델)
- **kosmos-2** -- [Kosmos2Model](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2Model) (Kosmos2Config 모델)
- **kosmos-2.5** -- [Kosmos2_5Model](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5Model) (Kosmos2_5Config 모델)
- **kyutai_speech_to_text** -- [KyutaiSpeechToTextModel](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextModel) (KyutaiSpeechToTextConfig 모델)
- **laguna** -- [LagunaModel](/docs/transformers/v5.23.0/ko/model_doc/laguna#transformers.LagunaModel) (LagunaConfig 모델)
- **lasr_ctc** -- [LasrForCTC](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrForCTC) (LasrCTCConfig 모델)
- **lasr_encoder** -- [LasrEncoder](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrEncoder) (LasrEncoderConfig 모델)
- **layoutlm** -- [LayoutLMModel](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMModel) (LayoutLMConfig 모델)
- **layoutlmv2** -- [LayoutLMv2Model](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2Model) (LayoutLMv2Config 모델)
- **layoutlmv3** -- [LayoutLMv3Model](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3Model) (LayoutLMv3Config 모델)
- **led** -- [LEDModel](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.LEDModel) (LEDConfig 모델)
- **levit** -- [LevitModel](/docs/transformers/v5.23.0/ko/model_doc/levit#transformers.LevitModel) (LevitConfig 모델)
- **lfm2** -- [Lfm2Model](/docs/transformers/v5.23.0/ko/model_doc/lfm2#transformers.Lfm2Model) (Lfm2Config 모델)
- **lfm2_moe** -- [Lfm2MoeModel](/docs/transformers/v5.23.0/ko/model_doc/lfm2_moe#transformers.Lfm2MoeModel) (Lfm2MoeConfig 모델)
- **lfm2_vl** -- [Lfm2VlModel](/docs/transformers/v5.23.0/ko/model_doc/lfm2_vl#transformers.Lfm2VlModel) (Lfm2VlConfig 모델)
- **lfm3** -- [Lfm3Model](/docs/transformers/v5.23.0/ko/model_doc/lfm3#transformers.Lfm3Model) (Lfm3Config 모델)
- **lfm3_vl** -- [Lfm3VlModel](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlModel) (Lfm3VlConfig 모델)
- **lfm3_vl_vision** -- [Lfm3VlVisionModel](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlVisionModel) (Lfm3VlVisionConfig 모델)
- **lightglue** -- [LightGlueForKeypointMatching](/docs/transformers/v5.23.0/ko/model_doc/lightglue#transformers.LightGlueForKeypointMatching) (LightGlueConfig 모델)
- **lighton_ocr** -- [LightOnOcrModel](/docs/transformers/v5.23.0/ko/model_doc/lighton_ocr#transformers.LightOnOcrModel) (LightOnOcrConfig 모델)
- **lilt** -- [LiltModel](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltModel) (LiltConfig 모델)
- **llama** -- [LlamaModel](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaModel) (LlamaConfig 모델)
- **llama4** -- [Llama4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4ForConditionalGeneration) (Llama4Config 모델)
- **llama4_text** -- [Llama4TextModel](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4TextModel) (Llama4TextConfig 모델)
- **llava** -- [LlavaModel](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaModel) (LlavaConfig 모델)
- **llava_next** -- [LlavaNextModel](/docs/transformers/v5.23.0/ko/model_doc/llava_next#transformers.LlavaNextModel) (LlavaNextConfig 모델)
- **llava_next_video** -- [LlavaNextVideoModel](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoModel) (LlavaNextVideoConfig 모델)
- **llava_onevision** -- [LlavaOnevisionModel](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionModel) (LlavaOnevisionConfig 모델)
- **longcat_flash** -- [LongcatFlashModel](/docs/transformers/v5.23.0/ko/model_doc/longcat_flash#transformers.LongcatFlashModel) (LongcatFlashConfig 모델)
- **longformer** -- [LongformerModel](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerModel) (LongformerConfig 모델)
- **longt5** -- [LongT5Model](/docs/transformers/v5.23.0/ko/model_doc/longt5#transformers.LongT5Model) (LongT5Config 모델)
- **luke** -- [LukeModel](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeModel) (LukeConfig 모델)
- **lw_detr** -- [LwDetrModel](/docs/transformers/v5.23.0/ko/model_doc/lw_detr#transformers.LwDetrModel) (LwDetrConfig 모델)
- **lxmert** -- [LxmertModel](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.LxmertModel) (LxmertConfig 모델)
- **m2m_100** -- [M2M100Model](/docs/transformers/v5.23.0/ko/model_doc/m2m_100#transformers.M2M100Model) (M2M100Config 모델)
- **mamba** -- [MambaModel](/docs/transformers/v5.23.0/ko/model_doc/mamba#transformers.MambaModel) (MambaConfig 모델)
- **mamba2** -- [Mamba2Model](/docs/transformers/v5.23.0/ko/model_doc/mamba2#transformers.Mamba2Model) (Mamba2Config 모델)
- **marian** -- [MarianModel](/docs/transformers/v5.23.0/ko/model_doc/marian#transformers.MarianModel) (MarianConfig 모델)
- **markuplm** -- [MarkupLMModel](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMModel) (MarkupLMConfig 모델)
- **mask2former** -- [Mask2FormerModel](/docs/transformers/v5.23.0/ko/model_doc/mask2former#transformers.Mask2FormerModel) (Mask2FormerConfig 모델)
- **maskformer** -- [MaskFormerModel](/docs/transformers/v5.23.0/ko/model_doc/maskformer#transformers.MaskFormerModel) (MaskFormerConfig 모델)
- **maskformer-swin** -- `MaskFormerSwinModel` (MaskFormerSwinConfig 모델)
- **mbart** -- [MBartModel](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartModel) (MBartConfig 모델)
- **megatron-bert** -- [MegatronBertModel](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertModel) (MegatronBertConfig 모델)
- **mellum** -- [MellumModel](/docs/transformers/v5.23.0/ko/model_doc/mellum#transformers.MellumModel) (MellumConfig 모델)
- **metaclip_2** -- [MetaClip2Model](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2Model) (MetaClip2Config 모델)
- **mgp-str** -- [MgpstrForSceneTextRecognition](/docs/transformers/v5.23.0/ko/model_doc/mgp-str#transformers.MgpstrForSceneTextRecognition) (MgpstrConfig 모델)
- **mimi** -- [MimiModel](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiModel) (MimiConfig 모델)
- **mimi_decoder** -- [MimiDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiDecoderModel) (MimiDecoderConfig 모델)
- **mimi_encoder** -- [MimiEncoderModel](/docs/transformers/v5.23.0/ko/model_doc/mimi#transformers.MimiEncoderModel) (MimiEncoderConfig 모델)
- **mimo_v2_flash** -- [MiMoV2FlashModel](/docs/transformers/v5.23.0/ko/model_doc/mimo_v2_flash#transformers.MiMoV2FlashModel) (MiMoV2FlashConfig 모델)
- **minicpm3** -- [MiniCPM3Model](/docs/transformers/v5.23.0/ko/model_doc/minicpm3#transformers.MiniCPM3Model) (MiniCPM3Config 모델)
- **minicpm_ocr** -- [MiniCPMOcrModel](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrModel) (MiniCPMOcrConfig 모델)
- **minicpm_ocr_vision** -- [MiniCPMOcrVisionModel](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrVisionModel) (MiniCPMOcrVisionConfig 모델)
- **minicpmv4_6** -- [MiniCPMV4_6Model](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6Model) (MiniCPMV4_6Config 모델)
- **minicpmv4_7** -- [MiniCPMV4_7Model](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7Model) (MiniCPMV4_7Config 모델)
- **minicpmv4_7_vision** -- [MiniCPMV4_7VisionModel](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7VisionModel) (MiniCPMV4_7VisionConfig 모델)
- **minimax** -- [MiniMaxModel](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxModel) (MiniMaxConfig 모델)
- **minimax_m2** -- [MiniMaxM2Model](/docs/transformers/v5.23.0/ko/model_doc/minimax_m2#transformers.MiniMaxM2Model) (MiniMaxM2Config 모델)
- **minimax_m3_vl** -- [MiniMaxM3VLModel](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLModel) (MiniMaxM3VLConfig 모델)
- **minimax_m3_vl_text** -- [MiniMaxM3VLTextModel](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLTextModel) (MiniMaxM3VLTextConfig 모델)
- **ministral** -- [MinistralModel](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralModel) (MinistralConfig 모델)
- **ministral3** -- [Ministral3Model](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3Model) (Ministral3Config 모델)
- **mistral** -- [MistralModel](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralModel) (MistralConfig 모델)
- **mistral3** -- [Mistral3Model](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3Model) (Mistral3Config 모델)
- **mistral4** -- [Mistral4Model](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4Model) (Mistral4Config 모델)
- **mistral_large4** -- [MistralLarge4Model](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4Model) (MistralLarge4Config 모델)
- **mistral_large4_text** -- [MistralLarge4TextModel](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4TextModel) (MistralLarge4TextConfig 모델)
- **mixtral** -- [MixtralModel](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralModel) (MixtralConfig 모델)
- **mlcd** -- [MLCDVisionModel](/docs/transformers/v5.23.0/ko/model_doc/mlcd#transformers.MLCDVisionModel) (MLCDVisionConfig 모델)
- **mlcd_vision_model** -- [MLCDVisionModel](/docs/transformers/v5.23.0/ko/model_doc/mlcd#transformers.MLCDVisionModel) (MLCDVisionConfig 모델)
- **mllama** -- [MllamaModel](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaModel) (MllamaConfig 모델)
- **mm-grounding-dino** -- [MMGroundingDinoModel](/docs/transformers/v5.23.0/ko/model_doc/mm-grounding-dino#transformers.MMGroundingDinoModel) (MMGroundingDinoConfig 모델)
- **mobilebert** -- [MobileBertModel](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertModel) (MobileBertConfig 모델)
- **mobilenet_v1** -- [MobileNetV1Model](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v1#transformers.MobileNetV1Model) (MobileNetV1Config 모델)
- **mobilenet_v2** -- [MobileNetV2Model](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v2#transformers.MobileNetV2Model) (MobileNetV2Config 모델)
- **mobilevit** -- [MobileViTModel](/docs/transformers/v5.23.0/ko/model_doc/mobilevit#transformers.MobileViTModel) (MobileViTConfig 모델)
- **mobilevitv2** -- [MobileViTV2Model](/docs/transformers/v5.23.0/ko/model_doc/mobilevitv2#transformers.MobileViTV2Model) (MobileViTV2Config 모델)
- **modernbert** -- [ModernBertModel](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertModel) (ModernBertConfig 모델)
- **modernbert-decoder** -- [ModernBertDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/modernbert-decoder#transformers.ModernBertDecoderModel) (ModernBertDecoderConfig 모델)
- **modernvbert** -- [ModernVBertModel](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertModel) (ModernVBertConfig 모델)
- **molmo2** -- [Molmo2Model](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2Model) (Molmo2Config 모델)
- **molmo2_text** -- [Molmo2TextModel](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2TextModel) (Molmo2TextConfig 모델)
- **moonshine** -- [MoonshineModel](/docs/transformers/v5.23.0/ko/model_doc/moonshine#transformers.MoonshineModel) (MoonshineConfig 모델)
- **moonshine_streaming** -- [MoonshineStreamingModel](/docs/transformers/v5.23.0/ko/model_doc/moonshine_streaming#transformers.MoonshineStreamingModel) (MoonshineStreamingConfig 모델)
- **moshi** -- [MoshiModel](/docs/transformers/v5.23.0/ko/model_doc/moshi#transformers.MoshiModel) (MoshiConfig 모델)
- **moss_transcribe_diarize** -- [MossTranscribeDiarizeModel](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeModel) (MossTranscribeDiarizeConfig 모델)
- **moss_transcribe_diarize_audio** -- [MossTranscribeDiarizeAudioModel](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeAudioModel) (MossTranscribeDiarizeAudioConfig 모델)
- **mpnet** -- [MPNetModel](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetModel) (MPNetConfig 모델)
- **mpt** -- [MptModel](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptModel) (MptConfig 모델)
- **mra** -- [MraModel](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraModel) (MraConfig 모델)
- **mt5** -- [MT5Model](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5Model) (MT5Config 모델)
- **muse_glimmer** -- [MuseGlimmerModel](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerModel) (MuseGlimmerConfig 모델)
- **muse_glimmer_assistant** -- [MuseGlimmerAssistantModel](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer_assistant#transformers.MuseGlimmerAssistantModel) (MuseGlimmerAssistantConfig 모델)
- **muse_glimmer_text** -- [MuseGlimmerTextModel](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerTextModel) (MuseGlimmerTextConfig 모델)
- **muse_glimmer_vision** -- [MuseGlimmerVisionModel](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerVisionModel) (MuseGlimmerVisionConfig 모델)
- **muse_spark** -- [MuseSparkModel](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkModel) (MuseSparkConfig 모델)
- **muse_spark_audio** -- [MuseSparkAudioModel](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkAudioModel) (MuseSparkAudioConfig 모델)
- **muse_spark_text** -- [MuseSparkTextModel](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkTextModel) (MuseSparkTextConfig 모델)
- **muse_spark_vision** -- [MuseSparkVisionModel](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkVisionModel) (MuseSparkVisionConfig 모델)
- **musicflamingo** -- [MusicFlamingoModel](/docs/transformers/v5.23.0/ko/model_doc/musicflamingo#transformers.MusicFlamingoModel) (MusicFlamingoConfig 모델)
- **musicgen** -- [MusicgenModel](/docs/transformers/v5.23.0/ko/model_doc/musicgen#transformers.MusicgenModel) (MusicgenConfig 모델)
- **musicgen_melody** -- [MusicgenMelodyModel](/docs/transformers/v5.23.0/ko/model_doc/musicgen_melody#transformers.MusicgenMelodyModel) (MusicgenMelodyConfig 모델)
- **mvp** -- [MvpModel](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpModel) (MvpConfig 모델)
- **nanochat** -- [NanoChatModel](/docs/transformers/v5.23.0/ko/model_doc/nanochat#transformers.NanoChatModel) (NanoChatConfig 모델)
- **nemo_speechlm** -- [Nemotron3_5TranscribeModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeModel) (Nemotron3_5TranscribeConfig 모델)
- **nemotron** -- [NemotronModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronModel) (NemotronConfig 모델)
- **nemotron3_5_asr** -- [Nemotron3_5AsrForRNNT](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_asr#transformers.Nemotron3_5AsrForRNNT) (Nemotron3_5AsrConfig 모델)
- **nemotron3_5_transcribe_asr_audio** -- [Nemotron3_5TranscribeAsrAudioModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeAsrAudioModel) (Nemotron3_5TranscribeAsrAudioConfig 모델)
- **nemotron3_5_transcribe_audio** -- [Nemotron3_5TranscribeAudioModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeAudioModel) (Nemotron3_5TranscribeAudioConfig 모델)
- **nemotron3_5_transcribe_word_timestamp** -- [Nemotron3_5TranscribeWordTimestampModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeWordTimestampModel) (Nemotron3_5TranscribeWordTimestampConfig 모델)
- **nemotron3_diarization** -- [Nemotron3DiarizationModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationModel) (Nemotron3DiarizationConfig 모델)
- **nemotron3_diarization_audio** -- [Nemotron3DiarizationAudioModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationAudioModel) (Nemotron3DiarizationAudioConfig 모델)
- **nemotron_asr_streaming** -- [NemotronAsrStreamingForRNNT](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingForRNNT) (NemotronAsrStreamingConfig 모델)
- **nemotron_asr_streaming_encoder** -- [NemotronAsrStreamingEncoder](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingEncoder) (NemotronAsrStreamingEncoderConfig 모델)
- **nemotron_h** -- [NemotronHModel](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h#transformers.NemotronHModel) (NemotronHConfig 모델)
- **neomme** -- [NeoMMEModel](/docs/transformers/v5.23.0/ko/model_doc/neomme#transformers.NeoMMEModel) (NeoMMEConfig 모델)
- **neucodec** -- [NeuCodecModel](/docs/transformers/v5.23.0/ko/model_doc/neucodec#transformers.NeuCodecModel) (NeuCodecConfig 모델)
- **nllb-moe** -- [NllbMoeModel](/docs/transformers/v5.23.0/ko/model_doc/nllb-moe#transformers.NllbMoeModel) (NllbMoeConfig 모델)
- **nomic_bert** -- [NomicBertModel](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertModel) (NomicBertConfig 모델)
- **nystromformer** -- [NystromformerModel](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerModel) (NystromformerConfig 모델)
- **olmo** -- [OlmoModel](/docs/transformers/v5.23.0/ko/model_doc/olmo#transformers.OlmoModel) (OlmoConfig 모델)
- **olmo2** -- [Olmo2Model](/docs/transformers/v5.23.0/ko/model_doc/olmo2#transformers.Olmo2Model) (Olmo2Config 모델)
- **olmo3** -- [Olmo3Model](/docs/transformers/v5.23.0/ko/model_doc/olmo3#transformers.Olmo3Model) (Olmo3Config 모델)
- **olmo_hybrid** -- [OlmoHybridModel](/docs/transformers/v5.23.0/ko/model_doc/olmo_hybrid#transformers.OlmoHybridModel) (OlmoHybridConfig 모델)
- **olmoe** -- [OlmoeModel](/docs/transformers/v5.23.0/ko/model_doc/olmoe#transformers.OlmoeModel) (OlmoeConfig 모델)
- **omdet-turbo** -- [OmDetTurboForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/omdet-turbo#transformers.OmDetTurboForObjectDetection) (OmDetTurboConfig 모델)
- **oneformer** -- [OneFormerModel](/docs/transformers/v5.23.0/ko/model_doc/oneformer#transformers.OneFormerModel) (OneFormerConfig 모델)
- **openai-gpt** -- [OpenAIGPTModel](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTModel) (OpenAIGPTConfig 모델)
- **openai_privacy_filter** -- [OpenAIPrivacyFilterModel](/docs/transformers/v5.23.0/ko/model_doc/openai_privacy_filter#transformers.OpenAIPrivacyFilterModel) (OpenAIPrivacyFilterConfig 모델)
- **opt** -- [OPTModel](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTModel) (OPTConfig 모델)
- **ovis2** -- [Ovis2Model](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2Model) (Ovis2Config 모델)
- **owlv2** -- [Owlv2Model](/docs/transformers/v5.23.0/ko/model_doc/owlv2#transformers.Owlv2Model) (Owlv2Config 모델)
- **owlvit** -- [OwlViTModel](/docs/transformers/v5.23.0/ko/model_doc/owlvit#transformers.OwlViTModel) (OwlViTConfig 모델)
- **paddleocr_vl** -- [PaddleOCRVLModel](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVLModel) (PaddleOCRVLConfig 모델)
- **paligemma** -- [PaliGemmaModel](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaModel) (PaliGemmaConfig 모델)
- **parakeet_ctc** -- [ParakeetForCTC](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForCTC) (ParakeetCTCConfig 모델)
- **parakeet_encoder** -- [ParakeetEncoder](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetEncoder) (ParakeetEncoderConfig 모델)
- **parakeet_rnnt** -- [ParakeetForRNNT](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForRNNT) (ParakeetRNNTConfig 모델)
- **parakeet_tdt** -- [ParakeetForTDT](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForTDT) (ParakeetTDTConfig 모델)
- **patchtsmixer** -- [PatchTSMixerModel](/docs/transformers/v5.23.0/ko/model_doc/patchtsmixer#transformers.PatchTSMixerModel) (PatchTSMixerConfig 모델)
- **patchtst** -- [PatchTSTModel](/docs/transformers/v5.23.0/ko/model_doc/patchtst#transformers.PatchTSTModel) (PatchTSTConfig 모델)
- **pe_audio** -- [PeAudioModel](/docs/transformers/v5.23.0/ko/model_doc/pe_audio#transformers.PeAudioModel) (PeAudioConfig 모델)
- **pe_audio_encoder** -- [PeAudioEncoder](/docs/transformers/v5.23.0/ko/model_doc/pe_audio#transformers.PeAudioEncoder) (PeAudioEncoderConfig 모델)
- **pe_audio_video** -- [PeAudioVideoModel](/docs/transformers/v5.23.0/ko/model_doc/pe_audio_video#transformers.PeAudioVideoModel) (PeAudioVideoConfig 모델)
- **pe_audio_video_encoder** -- [PeAudioVideoEncoder](/docs/transformers/v5.23.0/ko/model_doc/pe_audio_video#transformers.PeAudioVideoEncoder) (PeAudioVideoEncoderConfig 모델)
- **pe_video** -- [PeVideoModel](/docs/transformers/v5.23.0/ko/model_doc/pe_video#transformers.PeVideoModel) (PeVideoConfig 모델)
- **pe_video_encoder** -- [PeVideoEncoder](/docs/transformers/v5.23.0/ko/model_doc/pe_video#transformers.PeVideoEncoder) (PeVideoEncoderConfig 모델)
- **pegasus** -- [PegasusModel](/docs/transformers/v5.23.0/ko/model_doc/pegasus#transformers.PegasusModel) (PegasusConfig 모델)
- **pegasus_x** -- [PegasusXModel](/docs/transformers/v5.23.0/ko/model_doc/pegasus_x#transformers.PegasusXModel) (PegasusXConfig 모델)
- **perceiver** -- [PerceiverModel](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverModel) (PerceiverConfig 모델)
- **perception_lm** -- [PerceptionLMModel](/docs/transformers/v5.23.0/ko/model_doc/perception_lm#transformers.PerceptionLMModel) (PerceptionLMConfig 모델)
- **persimmon** -- [PersimmonModel](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonModel) (PersimmonConfig 모델)
- **phi** -- [PhiModel](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiModel) (PhiConfig 모델)
- **phi3** -- [Phi3Model](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3Model) (Phi3Config 모델)
- **phi4_multimodal** -- [Phi4MultimodalModel](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalModel) (Phi4MultimodalConfig 모델)
- **phimoe** -- [PhimoeModel](/docs/transformers/v5.23.0/ko/model_doc/phimoe#transformers.PhimoeModel) (PhimoeConfig 모델)
- **pi0** -- [PI0Model](/docs/transformers/v5.23.0/ko/model_doc/pi0#transformers.PI0Model) (PI0Config 모델)
- **pixio** -- [PixioModel](/docs/transformers/v5.23.0/ko/model_doc/pixio#transformers.PixioModel) (PixioConfig 모델)
- **pixtral** -- [PixtralVisionModel](/docs/transformers/v5.23.0/ko/model_doc/pixtral#transformers.PixtralVisionModel) (PixtralVisionConfig 모델)
- **plbart** -- [PLBartModel](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartModel) (PLBartConfig 모델)
- **pocket_tts** -- [PocketTTSModel](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts#transformers.PocketTTSModel) (PocketTTSConfig 모델)
- **pocket_tts_audio_tokenizer** -- [PocketTTSAudioTokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerModel) (PocketTTSAudioTokenizerConfig 모델)
- **pocket_tts_audio_tokenizer_decoder** -- [PocketTTSAudioTokenizerDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerDecoderModel) (PocketTTSAudioTokenizerDecoderConfig 모델)
- **pocket_tts_audio_tokenizer_encoder** -- [PocketTTSAudioTokenizerEncoderModel](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerEncoderModel) (PocketTTSAudioTokenizerEncoderConfig 모델)
- **poolformer** -- [PoolFormerModel](/docs/transformers/v5.23.0/ko/model_doc/poolformer#transformers.PoolFormerModel) (PoolFormerConfig 모델)
- **pp_doclayout_v3** -- [PPDocLayoutV3Model](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v3#transformers.PPDocLayoutV3Model) (PPDocLayoutV3Config 모델)
- **pp_doclayout_v4** -- [PPDocLayoutV4Model](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v4#transformers.PPDocLayoutV4Model) (PPDocLayoutV4Config 모델)
- **pp_ocrv5_mobile_rec** -- [PPOCRV5MobileRecModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_mobile_rec#transformers.PPOCRV5MobileRecModel) (PPOCRV5MobileRecConfig 모델)
- **pp_ocrv5_server_rec** -- [PPOCRV5ServerRecModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_server_rec#transformers.PPOCRV5ServerRecModel) (PPOCRV5ServerRecConfig 모델)
- **pp_ocrv6_medium_det** -- [PPOCRV6MediumDetModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_medium_det#transformers.PPOCRV6MediumDetModel) (PPOCRV6MediumDetConfig 모델)
- **pp_ocrv6_small_det** -- [PPOCRV6SmallDetModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_det#transformers.PPOCRV6SmallDetModel) (PPOCRV6SmallDetConfig 모델)
- **pp_ocrv6_small_rec** -- [PPOCRV6SmallRecModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_rec#transformers.PPOCRV6SmallRecModel) (PPOCRV6SmallRecConfig 모델)
- **pp_ocrv6_tiny_rec** -- [PPOCRV6TinyRecModel](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_tiny_rec#transformers.PPOCRV6TinyRecModel) (PPOCRV6TinyRecConfig 모델)
- **prophetnet** -- [ProphetNetModel](/docs/transformers/v5.23.0/ko/model_doc/prophetnet#transformers.ProphetNetModel) (ProphetNetConfig 모델)
- **pvt** -- [PvtModel](/docs/transformers/v5.23.0/ko/model_doc/pvt#transformers.PvtModel) (PvtConfig 모델)
- **pvt_v2** -- [PvtV2Model](/docs/transformers/v5.23.0/ko/model_doc/pvt_v2#transformers.PvtV2Model) (PvtV2Config 모델)
- **qianfan_ocr** -- [QianfanOCRModel](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRModel) (QianfanOCRConfig 모델)
- **qianfan_ocr_vision** -- [QianfanOCRVisionModel](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRVisionModel) (QianfanOCRVisionConfig 모델)
- **qwen2** -- [Qwen2Model](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Model) (Qwen2Config 모델)
- **qwen2_5_omni_thinker** -- [Qwen2_5OmniThinkerModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniThinkerModel) (Qwen2_5OmniThinkerConfig 모델)
- **qwen2_5_vl** -- [Qwen2_5_VLModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLModel) (Qwen2_5_VLConfig 모델)
- **qwen2_5_vl_text** -- [Qwen2_5_VLTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLTextModel) (Qwen2_5_VLTextConfig 모델)
- **qwen2_5_vl_vision** -- [Qwen2_5_VisionTransformerPretrainedModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VisionTransformerPretrainedModel) (Qwen2_5_VLVisionConfig 모델)
- **qwen2_audio** -- [Qwen2AudioModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioModel) (Qwen2AudioConfig 모델)
- **qwen2_audio_encoder** -- [Qwen2AudioEncoder](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioEncoder) (Qwen2AudioEncoderConfig 모델)
- **qwen2_moe** -- [Qwen2MoeModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeModel) (Qwen2MoeConfig 모델)
- **qwen2_vl** -- [Qwen2VLModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLModel) (Qwen2VLConfig 모델)
- **qwen2_vl_text** -- [Qwen2VLTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLTextModel) (Qwen2VLTextConfig 모델)
- **qwen3** -- [Qwen3Model](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3Model) (Qwen3Config 모델)
- **qwen3_5** -- [Qwen3_5Model](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Model) (Qwen3_5Config 모델)
- **qwen3_5_moe** -- [Qwen3_5MoeModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeModel) (Qwen3_5MoeConfig 모델)
- **qwen3_5_moe_text** -- [Qwen3_5MoeTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeTextModel) (Qwen3_5MoeTextConfig 모델)
- **qwen3_5_moe_vision** -- [Qwen3_5MoeVisionModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeVisionModel) (Qwen3_5MoeVisionConfig 모델)
- **qwen3_5_text** -- [Qwen3_5TextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5TextModel) (Qwen3_5TextConfig 모델)
- **qwen3_5_vision** -- [Qwen3_5VisionModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5VisionModel) (Qwen3_5VisionConfig 모델)
- **qwen3_asr** -- [Qwen3ASRModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRModel) (Qwen3ASRConfig 모델)
- **qwen3_asr_encoder** -- [Qwen3ASREncoder](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASREncoder) (Qwen3ASREncoderConfig 모델)
- **qwen3_moe** -- [Qwen3MoeModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeModel) (Qwen3MoeConfig 모델)
- **qwen3_next** -- [Qwen3NextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextModel) (Qwen3NextConfig 모델)
- **qwen3_omni_moe_thinker** -- [Qwen3OmniMoeThinkerModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeThinkerModel) (Qwen3OmniMoeThinkerConfig 모델)
- **qwen3_vl** -- [Qwen3VLModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLModel) (Qwen3VLConfig 모델)
- **qwen3_vl_moe** -- [Qwen3VLMoeModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeModel) (Qwen3VLMoeConfig 모델)
- **qwen3_vl_moe_text** -- [Qwen3VLMoeTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeTextModel) (Qwen3VLMoeTextConfig 모델)
- **qwen3_vl_moe_vision** -- [Qwen3VLMoeVisionModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeVisionModel) (Qwen3VLMoeVisionConfig 모델)
- **qwen3_vl_text** -- [Qwen3VLTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLTextModel) (Qwen3VLTextConfig 모델)
- **qwen3_vl_vision** -- [Qwen3VLVisionModel](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLVisionModel) (Qwen3VLVisionConfig 모델)
- **qwen4_exp** -- [Qwen4ExpModel](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpModel) (Qwen4ExpConfig 모델)
- **qwen4_exp_text** -- [Qwen4ExpTextModel](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpTextModel) (Qwen4ExpTextConfig 모델)
- **qwen4_exp_vision** -- [Qwen4ExpVisionModel](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpVisionModel) (Qwen4ExpVisionConfig 모델)
- **radio** -- [RadioModel](/docs/transformers/v5.23.0/ko/model_doc/radio#transformers.RadioModel) (RadioConfig 모델)
- **recurrent_gemma** -- [RecurrentGemmaModel](/docs/transformers/v5.23.0/ko/model_doc/recurrent_gemma#transformers.RecurrentGemmaModel) (RecurrentGemmaConfig 모델)
- **reformer** -- [ReformerModel](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerModel) (ReformerConfig 모델)
- **regnet** -- [RegNetModel](/docs/transformers/v5.23.0/ko/model_doc/regnet#transformers.RegNetModel) (RegNetConfig 모델)
- **rembert** -- [RemBertModel](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertModel) (RemBertConfig 모델)
- **resnet** -- [ResNetModel](/docs/transformers/v5.23.0/ko/model_doc/resnet#transformers.ResNetModel) (ResNetConfig 모델)
- **rf_detr** -- [RfDetrModel](/docs/transformers/v5.23.0/ko/model_doc/rf_detr#transformers.RfDetrModel) (RfDetrConfig 모델)
- **roberta** -- [RobertaModel](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaModel) (RobertaConfig 모델)
- **roberta-prelayernorm** -- [RobertaPreLayerNormModel](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormModel) (RobertaPreLayerNormConfig 모델)
- **roc_bert** -- [RoCBertModel](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertModel) (RoCBertConfig 모델)
- **roformer** -- [RoFormerModel](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerModel) (RoFormerConfig 모델)
- **rt_detr** -- [RTDetrModel](/docs/transformers/v5.23.0/ko/model_doc/rt_detr#transformers.RTDetrModel) (RTDetrConfig 모델)
- **rt_detr_v2** -- [RTDetrV2Model](/docs/transformers/v5.23.0/ko/model_doc/rt_detr_v2#transformers.RTDetrV2Model) (RTDetrV2Config 모델)
- **rwkv** -- [RwkvModel](/docs/transformers/v5.23.0/ko/model_doc/rwkv#transformers.RwkvModel) (RwkvConfig 모델)
- **sam** -- [SamModel](/docs/transformers/v5.23.0/ko/model_doc/sam#transformers.SamModel) (SamConfig 모델)
- **sam2** -- [Sam2Model](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2Model) (Sam2Config 모델)
- **sam2_hiera_det_model** -- [Sam2HieraDetModel](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2HieraDetModel) (Sam2HieraDetConfig 모델)
- **sam2_video** -- [Sam2VideoModel](/docs/transformers/v5.23.0/ko/model_doc/sam2_video#transformers.Sam2VideoModel) (Sam2VideoConfig 모델)
- **sam2_vision_model** -- [Sam2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/sam2#transformers.Sam2VisionModel) (Sam2VisionConfig 모델)
- **sam3** -- [Sam3Model](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3Model) (Sam3Config 모델)
- **sam3_lite_text** -- [Sam3LiteTextModel](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextModel) (Sam3LiteTextConfig 모델)
- **sam3_lite_text_text_model** -- [Sam3LiteTextTextModel](/docs/transformers/v5.23.0/ko/model_doc/sam3_lite_text#transformers.Sam3LiteTextTextModel) (Sam3LiteTextTextConfig 모델)
- **sam3_tracker** -- [Sam3TrackerModel](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker#transformers.Sam3TrackerModel) (Sam3TrackerConfig 모델)
- **sam3_tracker_video** -- [Sam3TrackerVideoModel](/docs/transformers/v5.23.0/ko/model_doc/sam3_tracker_video#transformers.Sam3TrackerVideoModel) (Sam3TrackerVideoConfig 모델)
- **sam3_video** -- [Sam3VideoModel](/docs/transformers/v5.23.0/ko/model_doc/sam3_video#transformers.Sam3VideoModel) (Sam3VideoConfig 모델)
- **sam3_vision_model** -- [Sam3VisionModel](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3VisionModel) (Sam3VisionConfig 모델)
- **sam3_vit_model** -- [Sam3ViTModel](/docs/transformers/v5.23.0/ko/model_doc/sam3#transformers.Sam3ViTModel) (Sam3ViTConfig 모델)
- **sam_hq** -- [SamHQModel](/docs/transformers/v5.23.0/ko/model_doc/sam_hq#transformers.SamHQModel) (SamHQConfig 모델)
- **sam_hq_vision_model** -- [SamHQVisionModel](/docs/transformers/v5.23.0/ko/model_doc/sam_hq#transformers.SamHQVisionModel) (SamHQVisionConfig 모델)
- **sam_vision_model** -- [SamVisionModel](/docs/transformers/v5.23.0/ko/model_doc/sam#transformers.SamVisionModel) (SamVisionConfig 모델)
- **sapiens2** -- [Sapiens2Model](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2Model) (Sapiens2Config 모델)
- **seamless_m4t** -- [SeamlessM4TModel](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TModel) (SeamlessM4TConfig 모델)
- **seamless_m4t_v2** -- [SeamlessM4Tv2Model](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t_v2#transformers.SeamlessM4Tv2Model) (SeamlessM4Tv2Config 모델)
- **seed_oss** -- [SeedOssModel](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssModel) (SeedOssConfig 모델)
- **segformer** -- [SegformerModel](/docs/transformers/v5.23.0/ko/model_doc/segformer#transformers.SegformerModel) (SegformerConfig 모델)
- **seggpt** -- [SegGptModel](/docs/transformers/v5.23.0/ko/model_doc/seggpt#transformers.SegGptModel) (SegGptConfig 모델)
- **sew** -- [SEWModel](/docs/transformers/v5.23.0/ko/model_doc/sew#transformers.SEWModel) (SEWConfig 모델)
- **sew-d** -- [SEWDModel](/docs/transformers/v5.23.0/ko/model_doc/sew-d#transformers.SEWDModel) (SEWDConfig 모델)
- **siglip** -- [SiglipModel](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipModel) (SiglipConfig 모델)
- **siglip2** -- [Siglip2Model](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2Model) (Siglip2Config 모델)
- **siglip2_vision_model** -- [Siglip2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2VisionModel) (Siglip2VisionConfig 모델)
- **siglip_vision_model** -- [SiglipVisionModel](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipVisionModel) (SiglipVisionConfig 모델)
- **smollm3** -- [SmolLM3Model](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3Model) (SmolLM3Config 모델)
- **smolvlm** -- [SmolVLMModel](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMModel) (SmolVLMConfig 모델)
- **smolvlm_vision** -- [SmolVLMVisionTransformer](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMVisionTransformer) (SmolVLMVisionConfig 모델)
- **solar_open** -- [SolarOpenModel](/docs/transformers/v5.23.0/ko/model_doc/solar_open#transformers.SolarOpenModel) (SolarOpenConfig 모델)
- **speech_to_text** -- [Speech2TextModel](/docs/transformers/v5.23.0/ko/model_doc/speech_to_text#transformers.Speech2TextModel) (Speech2TextConfig 모델)
- **speecht5** -- [SpeechT5Model](/docs/transformers/v5.23.0/ko/model_doc/speecht5#transformers.SpeechT5Model) (SpeechT5Config 모델)
- **splinter** -- [SplinterModel](/docs/transformers/v5.23.0/ko/model_doc/splinter#transformers.SplinterModel) (SplinterConfig 모델)
- **squeezebert** -- [SqueezeBertModel](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertModel) (SqueezeBertConfig 모델)
- **stablelm** -- [StableLmModel](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmModel) (StableLmConfig 모델)
- **starcoder2** -- [Starcoder2Model](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2Model) (Starcoder2Config 모델)
- **step3p7** -- [Step3p7Model](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7Model) (Step3p7Config 모델)
- **swiftformer** -- [SwiftFormerModel](/docs/transformers/v5.23.0/ko/model_doc/swiftformer#transformers.SwiftFormerModel) (SwiftFormerConfig 모델)
- **swin** -- [SwinModel](/docs/transformers/v5.23.0/ko/model_doc/swin#transformers.SwinModel) (SwinConfig 모델)
- **swin2sr** -- [Swin2SRModel](/docs/transformers/v5.23.0/ko/model_doc/swin2sr#transformers.Swin2SRModel) (Swin2SRConfig 모델)
- **swinv2** -- [Swinv2Model](/docs/transformers/v5.23.0/ko/model_doc/swinv2#transformers.Swinv2Model) (Swinv2Config 모델)
- **switch_transformers** -- [SwitchTransformersModel](/docs/transformers/v5.23.0/ko/model_doc/switch_transformers#transformers.SwitchTransformersModel) (SwitchTransformersConfig 모델)
- **t5** -- [T5Model](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Model) (T5Config 모델)
- **t5gemma** -- [T5GemmaModel](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaModel) (T5GemmaConfig 모델)
- **t5gemma2** -- [T5Gemma2Model](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2Model) (T5Gemma2Config 모델)
- **t5gemma2_encoder** -- `T5Gemma2Encoder` (T5Gemma2EncoderConfig 모델)
- **table-transformer** -- [TableTransformerModel](/docs/transformers/v5.23.0/ko/model_doc/table-transformer#transformers.TableTransformerModel) (TableTransformerConfig 모델)
- **tapas** -- [TapasModel](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasModel) (TapasConfig 모델)
- **textnet** -- [TextNetModel](/docs/transformers/v5.23.0/ko/model_doc/textnet#transformers.TextNetModel) (TextNetConfig 모델)
- **time_series_transformer** -- [TimeSeriesTransformerModel](/docs/transformers/v5.23.0/ko/model_doc/time_series_transformer#transformers.TimeSeriesTransformerModel) (TimeSeriesTransformerConfig 모델)
- **timesfm** -- [TimesFmModel](/docs/transformers/v5.23.0/ko/model_doc/timesfm#transformers.TimesFmModel) (TimesFmConfig 모델)
- **timesfm2_5** -- [TimesFm2_5Model](/docs/transformers/v5.23.0/ko/model_doc/timesfm2_5#transformers.TimesFm2_5Model) (TimesFm2_5Config 모델)
- **timesformer** -- [TimesformerModel](/docs/transformers/v5.23.0/ko/model_doc/timesformer#transformers.TimesformerModel) (TimesformerConfig 모델)
- **timm_backbone** -- [TimmBackbone](/docs/transformers/v5.23.0/ko/main_classes/backbones#transformers.TimmBackbone) (TimmBackboneConfig 모델)
- **timm_wrapper** -- [TimmWrapperModel](/docs/transformers/v5.23.0/ko/model_doc/timm_wrapper#transformers.TimmWrapperModel) (TimmWrapperConfig 모델)
- **tipsv2** -- [Tipsv2Model](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2Model) (Tipsv2Config 모델)
- **tipsv2_text_model** -- [Tipsv2TextModel](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2TextModel) (Tipsv2TextConfig 모델)
- **tipsv2_vision_model** -- [Tipsv2VisionModel](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2VisionModel) (Tipsv2VisionConfig 모델)
- **tvp** -- [TvpModel](/docs/transformers/v5.23.0/ko/model_doc/tvp#transformers.TvpModel) (TvpConfig 모델)
- **udop** -- [UdopModel](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopModel) (UdopConfig 모델)
- **ultrabert** -- [UltrabertModel](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertModel) (UltrabertConfig 모델)
- **ultravbert** -- [UltraVBertModel](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertModel) (UltraVBertConfig 모델)
- **umt5** -- [UMT5Model](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5Model) (UMT5Config 모델)
- **unispeech** -- [UniSpeechModel](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechModel) (UniSpeechConfig 모델)
- **unispeech-sat** -- [UniSpeechSatModel](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatModel) (UniSpeechSatConfig 모델)
- **univnet** -- [UnivNetModel](/docs/transformers/v5.23.0/ko/model_doc/univnet#transformers.UnivNetModel) (UnivNetConfig 모델)
- **unlimited_ocr** -- [UnlimitedOcrModel](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrModel) (UnlimitedOcrConfig 모델)
- **uvdoc** -- [UVDocModel](/docs/transformers/v5.23.0/ko/model_doc/uvdoc#transformers.UVDocModel) (UVDocConfig 모델)
- **vaultgemma** -- [VaultGemmaModel](/docs/transformers/v5.23.0/ko/model_doc/vaultgemma#transformers.VaultGemmaModel) (VaultGemmaConfig 모델)
- **vibevoice** -- [VibeVoiceForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice#transformers.VibeVoiceForConditionalGeneration) (VibeVoiceConfig 모델)
- **vibevoice_acoustic_tokenizer** -- [VibeVoiceAcousticTokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerModel) (VibeVoiceAcousticTokenizerConfig 모델)
- **vibevoice_acoustic_tokenizer_decoder** -- [VibeVoiceAcousticTokenizerDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerDecoderModel) (VibeVoiceAcousticTokenizerDecoderConfig 모델)
- **vibevoice_acoustic_tokenizer_encoder** -- [VibeVoiceAcousticTokenizerEncoderModel](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerEncoderModel) (VibeVoiceAcousticTokenizerEncoderConfig 모델)
- **vibevoice_asr** -- [VibeVoiceAsrModel](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrModel) (VibeVoiceAsrConfig 모델)
- **video_llama_3** -- [VideoLlama3Model](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3Model) (VideoLlama3Config 모델)
- **video_llama_3_vision** -- [VideoLlama3VisionModel](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3VisionModel) (VideoLlama3VisionConfig 모델)
- **video_llava** -- [VideoLlavaModel](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaModel) (VideoLlavaConfig 모델)
- **videomae** -- [VideoMAEModel](/docs/transformers/v5.23.0/ko/model_doc/videomae#transformers.VideoMAEModel) (VideoMAEConfig 모델)
- **videoprism** -- [VideoPrismClipModel](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismClipModel) (VideoPrismConfig 모델)
- **videoprism_vision_model** -- [VideoPrismVisionModel](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismVisionModel) (VideoPrismVisionConfig 모델)
- **vilt** -- [ViltModel](/docs/transformers/v5.23.0/ko/model_doc/vilt#transformers.ViltModel) (ViltConfig 모델)
- **vipllava** -- [VipLlavaModel](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaModel) (VipLlavaConfig 모델)
- **vision-text-dual-encoder** -- [VisionTextDualEncoderModel](/docs/transformers/v5.23.0/ko/model_doc/vision-text-dual-encoder#transformers.VisionTextDualEncoderModel) (VisionTextDualEncoderConfig 모델)
- **visual_bert** -- [VisualBertModel](/docs/transformers/v5.23.0/ko/model_doc/visual_bert#transformers.VisualBertModel) (VisualBertConfig 모델)
- **vit** -- [ViTModel](/docs/transformers/v5.23.0/ko/model_doc/vit#transformers.ViTModel) (ViTConfig 모델)
- **vit_mae** -- [ViTMAEModel](/docs/transformers/v5.23.0/ko/model_doc/vit_mae#transformers.ViTMAEModel) (ViTMAEConfig 모델)
- **vit_msn** -- [ViTMSNModel](/docs/transformers/v5.23.0/ko/model_doc/vit_msn#transformers.ViTMSNModel) (ViTMSNConfig 모델)
- **vitdet** -- [VitDetModel](/docs/transformers/v5.23.0/ko/model_doc/vitdet#transformers.VitDetModel) (VitDetConfig 모델)
- **vits** -- [VitsModel](/docs/transformers/v5.23.0/ko/model_doc/vits#transformers.VitsModel) (VitsConfig 모델)
- **vivit** -- [VivitModel](/docs/transformers/v5.23.0/ko/model_doc/vivit#transformers.VivitModel) (VivitConfig 모델)
- **vjepa2** -- [VJEPA2Model](/docs/transformers/v5.23.0/ko/model_doc/vjepa2#transformers.VJEPA2Model) (VJEPA2Config 모델)
- **voxtral** -- [VoxtralModel](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralModel) (VoxtralConfig 모델)
- **voxtral_encoder** -- [VoxtralEncoder](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralEncoder) (VoxtralEncoderConfig 모델)
- **voxtral_realtime** -- [VoxtralRealtimeModel](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeModel) (VoxtralRealtimeConfig 모델)
- **voxtral_realtime_encoder** -- [VoxtralRealtimeEncoder](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeEncoder) (VoxtralRealtimeEncoderConfig 모델)
- **voxtral_realtime_text** -- `VoxtralRealtimeTextModel` (VoxtralRealtimeTextConfig 모델)
- **wav2vec2** -- [Wav2Vec2Model](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Model) (Wav2Vec2Config 모델)
- **wav2vec2-bert** -- [Wav2Vec2BertModel](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertModel) (Wav2Vec2BertConfig 모델)
- **wav2vec2-conformer** -- [Wav2Vec2ConformerModel](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerModel) (Wav2Vec2ConformerConfig 모델)
- **wavlm** -- [WavLMModel](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMModel) (WavLMConfig 모델)
- **whisper** -- [WhisperModel](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperModel) (WhisperConfig 모델)
- **xclip** -- [XCLIPModel](/docs/transformers/v5.23.0/ko/model_doc/xclip#transformers.XCLIPModel) (XCLIPConfig 모델)
- **xcodec** -- [XcodecModel](/docs/transformers/v5.23.0/ko/model_doc/xcodec#transformers.XcodecModel) (XcodecConfig 모델)
- **xcodec2** -- [Xcodec2Model](/docs/transformers/v5.23.0/ko/model_doc/xcodec2#transformers.Xcodec2Model) (Xcodec2Config 모델)
- **xglm** -- [XGLMModel](/docs/transformers/v5.23.0/ko/model_doc/xglm#transformers.XGLMModel) (XGLMConfig 모델)
- **xlm** -- [XLMModel](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMModel) (XLMConfig 모델)
- **xlm-roberta** -- [XLMRobertaModel](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaModel) (XLMRobertaConfig 모델)
- **xlm-roberta-xl** -- [XLMRobertaXLModel](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLModel) (XLMRobertaXLConfig 모델)
- **xlnet** -- [XLNetModel](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetModel) (XLNetConfig 모델)
- **xlstm** -- [xLSTMModel](/docs/transformers/v5.23.0/ko/model_doc/xlstm#transformers.xLSTMModel) (xLSTMConfig 모델)
- **xmod** -- [XmodModel](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodModel) (XmodConfig 모델)
- **yolos** -- [YolosModel](/docs/transformers/v5.23.0/ko/model_doc/yolos#transformers.YolosModel) (YolosConfig 모델)
- **yoso** -- [YosoModel](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoModel) (YosoConfig 모델)
- **youtu** -- [YoutuModel](/docs/transformers/v5.23.0/ko/model_doc/youtu#transformers.YoutuModel) (YoutuConfig 모델)
- **zamba** -- [ZambaModel](/docs/transformers/v5.23.0/ko/model_doc/zamba#transformers.ZambaModel) (ZambaConfig 모델)
- **zamba2** -- [Zamba2Model](/docs/transformers/v5.23.0/ko/model_doc/zamba2#transformers.Zamba2Model) (Zamba2Config 모델)
- **zaya** -- [ZayaModel](/docs/transformers/v5.23.0/ko/model_doc/zaya#transformers.ZayaModel) (ZayaConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModel

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModel.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModel.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

## 범용 사전 학습 클래스

다음과 같은 auto 클래스를 사용하면 사전 학습 헤드가 있는 모델을 인스턴스화할 수 있습니다.

### AutoModelForPreTraining[[transformers.AutoModelForPreTraining]]

#### transformers.AutoModelForPreTraining[[transformers.AutoModelForPreTraining]]

```python
transformers.AutoModelForPreTraining(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2340)

이것은 범용 모델 클래스로, [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때 라이브러리의 모델 클래스 중 하나로(프리트레이닝 헤드를 포함하여) 인스턴스화됩니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForPreTraining.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : The model class to instantiate is selected based on the configuration class:  - [AlbertConfig](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertConfig) configuration class: [AlbertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForPreTraining) (AlbertConfig model) - [AudioFlamingo3Config](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3Config) configuration class: [AudioFlamingo3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3ForConditionalGeneration) (AudioFlamingo3Config model) - [BartConfig](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartConfig) configuration class: [BartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForConditionalGeneration) (BartConfig model) - [BertConfig](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertConfig) configuration class: [BertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForPreTraining) (BertConfig model) - [BigBirdConfig](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdConfig) configuration class: [BigBirdForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForPreTraining) (BigBirdConfig model) - [BloomConfig](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomConfig) configuration class: [BloomForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomForCausalLM) (BloomConfig model) - [CTRLConfig](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLConfig) configuration class: [CTRLLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLLMHeadModel) (CTRLConfig model) - [CamembertConfig](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertConfig) configuration class: [CamembertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForMaskedLM) (CamembertConfig model) - [ColModernVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/colmodernvbert#transformers.ColModernVBertConfig) configuration class: [ColModernVBertForRetrieval](/docs/transformers/v5.23.0/ko/model_doc/colmodernvbert#transformers.ColModernVBertForRetrieval) (ColModernVBertConfig model) - [ColPaliConfig](/docs/transformers/v5.23.0/ko/model_doc/colpali#transformers.ColPaliConfig) configuration class: [ColPaliForRetrieval](/docs/transformers/v5.23.0/ko/model_doc/colpali#transformers.ColPaliForRetrieval) (ColPaliConfig model) - [ColQwen2Config](/docs/transformers/v5.23.0/ko/model_doc/colqwen2#transformers.ColQwen2Config) configuration class: [ColQwen2ForRetrieval](/docs/transformers/v5.23.0/ko/model_doc/colqwen2#transformers.ColQwen2ForRetrieval) (ColQwen2Config model) - [Data2VecTextConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextConfig) configuration class: [Data2VecTextForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForMaskedLM) (Data2VecTextConfig model) - [DebertaConfig](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaConfig) configuration class: [DebertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaForMaskedLM) (DebertaConfig model) - [DebertaV2Config](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2Config) configuration class: [DebertaV2ForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForMaskedLM) (DebertaV2Config model) - [DistilBertConfig](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertConfig) configuration class: [DistilBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForMaskedLM) (DistilBertConfig model) - [ElectraConfig](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraConfig) configuration class: [ElectraForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForPreTraining) (ElectraConfig model) - [ErnieConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieConfig) configuration class: [ErnieForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForPreTraining) (ErnieConfig model) - [EvollaConfig](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaConfig) configuration class: [EvollaForProteinText2Text](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaForProteinText2Text) (EvollaConfig model) - [Exaone4Config](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4Config) configuration class: [Exaone4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4ForCausalLM) (Exaone4Config model) - [ExaoneMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/exaone_moe#transformers.ExaoneMoeConfig) configuration class: [ExaoneMoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/exaone_moe#transformers.ExaoneMoeForCausalLM) (ExaoneMoeConfig model) - [FNetConfig](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetConfig) configuration class: [FNetForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForPreTraining) (FNetConfig model) - [FSMTConfig](/docs/transformers/v5.23.0/ko/model_doc/fsmt#transformers.FSMTConfig) configuration class: [FSMTForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fsmt#transformers.FSMTForConditionalGeneration) (FSMTConfig model) - [FalconMambaConfig](/docs/transformers/v5.23.0/ko/model_doc/falcon_mamba#transformers.FalconMambaConfig) configuration class: [FalconMambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/falcon_mamba#transformers.FalconMambaForCausalLM) (FalconMambaConfig model) - [FlaubertConfig](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertConfig) configuration class: [FlaubertWithLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertWithLMHeadModel) (FlaubertConfig model) - [FlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaConfig) configuration class: [FlavaForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaForPreTraining) (FlavaConfig model) - [Florence2Config](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2Config) configuration class: [Florence2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2ForConditionalGeneration) (Florence2Config model) - [FunAsrNanoConfig](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoConfig) configuration class: [FunAsrNanoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoForConditionalGeneration) (FunAsrNanoConfig model) - [FunnelConfig](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelConfig) configuration class: [FunnelForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForPreTraining) (FunnelConfig model) - [GPT2Config](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Config) configuration class: [GPT2LMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2LMHeadModel) (GPT2Config model) - [GPTBigCodeConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeConfig) configuration class: [GPTBigCodeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeForCausalLM) (GPTBigCodeConfig model) - [Gemma3Config](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Config) configuration class: [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (Gemma3Config model) - [Gemma4Config](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4Config) configuration class: [Gemma4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4ForConditionalGeneration) (Gemma4Config model) - [Gemma4UnifiedConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedConfig) configuration class: [Gemma4UnifiedForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedForConditionalGeneration) (Gemma4UnifiedConfig model) - [GlmAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrConfig) configuration class: [GlmAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrForConditionalGeneration) (GlmAsrConfig model) - [HieraConfig](/docs/transformers/v5.23.0/ko/model_doc/hiera#transformers.HieraConfig) configuration class: [HieraForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/hiera#transformers.HieraForPreTraining) (HieraConfig model) - [IBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertConfig) configuration class: [IBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForMaskedLM) (IBertConfig model) - [Idefics2Config](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2Config) configuration class: [Idefics2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2ForConditionalGeneration) (Idefics2Config model) - [Idefics3Config](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3Config) configuration class: [Idefics3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3ForConditionalGeneration) (Idefics3Config model) - [IdeficsConfig](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsConfig) configuration class: [IdeficsForVisionText2Text](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsForVisionText2Text) (IdeficsConfig model) - [Inkling2Config](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2Config) configuration class: [Inkling2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2ForConditionalGeneration) (Inkling2Config model) - [InklingConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingConfig) configuration class: [InklingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingForConditionalGeneration) (InklingConfig model) - [JanusConfig](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusConfig) configuration class: [JanusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusForConditionalGeneration) (JanusConfig model) - [LayoutLMConfig](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMConfig) configuration class: [LayoutLMForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMForMaskedLM) (LayoutLMConfig model) - [LlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaConfig) configuration class: [LlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaForConditionalGeneration) (LlavaConfig model) - [LlavaNextConfig](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextConfig) configuration class: [LlavaNextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextForConditionalGeneration) (LlavaNextConfig model) - [LlavaNextVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoConfig) configuration class: [LlavaNextVideoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoForConditionalGeneration) (LlavaNextVideoConfig model) - [LlavaOnevisionConfig](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionConfig) configuration class: [LlavaOnevisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionForConditionalGeneration) (LlavaOnevisionConfig model) - [LongformerConfig](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerConfig) configuration class: [LongformerForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForMaskedLM) (LongformerConfig model) - [LukeConfig](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeConfig) configuration class: [LukeForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForMaskedLM) (LukeConfig model) - [LxmertConfig](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.LxmertConfig) configuration class: [LxmertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.LxmertForPreTraining) (LxmertConfig model) - [MPNetConfig](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetConfig) configuration class: [MPNetForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForMaskedLM) (MPNetConfig model) - [Mamba2Config](/docs/transformers/v5.23.0/ko/model_doc/mamba2#transformers.Mamba2Config) configuration class: [Mamba2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mamba2#transformers.Mamba2ForCausalLM) (Mamba2Config model) - [MambaConfig](/docs/transformers/v5.23.0/ko/model_doc/mamba#transformers.MambaConfig) configuration class: [MambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mamba#transformers.MambaForCausalLM) (MambaConfig model) - [MegatronBertConfig](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertConfig) configuration class: [MegatronBertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForPreTraining) (MegatronBertConfig model) - [Mistral3Config](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3Config) configuration class: [Mistral3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3ForConditionalGeneration) (Mistral3Config model) - [Mistral4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4Config) configuration class: [Mistral4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4ForCausalLM) (Mistral4Config model) - [MistralLarge4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4Config) configuration class: [MistralLarge4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4ForConditionalGeneration) (MistralLarge4Config model) - [MllamaConfig](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaConfig) configuration class: [MllamaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaForConditionalGeneration) (MllamaConfig model) - [MobileBertConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertConfig) configuration class: [MobileBertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForPreTraining) (MobileBertConfig model) - [MossTranscribeDiarizeConfig](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeConfig) configuration class: [MossTranscribeDiarizeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeForConditionalGeneration) (MossTranscribeDiarizeConfig model) - [MptConfig](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptConfig) configuration class: [MptForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptForCausalLM) (MptConfig model) - [MraConfig](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraConfig) configuration class: [MraForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForMaskedLM) (MraConfig model) - [MusicFlamingoConfig](/docs/transformers/v5.23.0/ko/model_doc/musicflamingo#transformers.MusicFlamingoConfig) configuration class: [MusicFlamingoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/musicflamingo#transformers.MusicFlamingoForConditionalGeneration) (MusicFlamingoConfig model) - [MvpConfig](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpConfig) configuration class: [MvpForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForConditionalGeneration) (MvpConfig model) - [NanoChatConfig](/docs/transformers/v5.23.0/ko/model_doc/nanochat#transformers.NanoChatConfig) configuration class: [NanoChatForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/nanochat#transformers.NanoChatForCausalLM) (NanoChatConfig model) - [Nemotron3_5TranscribeConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeConfig) configuration class: [Nemotron3_5TranscribeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeForConditionalGeneration) (Nemotron3_5TranscribeConfig model) - [NllbMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/nllb-moe#transformers.NllbMoeConfig) configuration class: [NllbMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/nllb-moe#transformers.NllbMoeForConditionalGeneration) (NllbMoeConfig model) - [OpenAIGPTConfig](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTConfig) configuration class: [OpenAIGPTLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTLMHeadModel) (OpenAIGPTConfig model) - [PaliGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaConfig) configuration class: [PaliGemmaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaForConditionalGeneration) (PaliGemmaConfig model) - [Qwen2AudioConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioConfig) configuration class: [Qwen2AudioForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioForConditionalGeneration) (Qwen2AudioConfig model) - [Qwen3ASRConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRConfig) configuration class: [Qwen3ASRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRForConditionalGeneration) (Qwen3ASRConfig model) - [RoCBertConfig](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertConfig) configuration class: [RoCBertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForPreTraining) (RoCBertConfig model) - [RobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaConfig) configuration class: [RobertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForMaskedLM) (RobertaConfig model) - [RobertaPreLayerNormConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormConfig) configuration class: [RobertaPreLayerNormForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForMaskedLM) (RobertaPreLayerNormConfig model) - [RwkvConfig](/docs/transformers/v5.23.0/ko/model_doc/rwkv#transformers.RwkvConfig) configuration class: [RwkvForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/rwkv#transformers.RwkvForCausalLM) (RwkvConfig model) - [SplinterConfig](/docs/transformers/v5.23.0/ko/model_doc/splinter#transformers.SplinterConfig) configuration class: [SplinterForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/splinter#transformers.SplinterForPreTraining) (SplinterConfig model) - [SqueezeBertConfig](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertConfig) configuration class: [SqueezeBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForMaskedLM) (SqueezeBertConfig model) - [SwitchTransformersConfig](/docs/transformers/v5.23.0/ko/model_doc/switch_transformers#transformers.SwitchTransformersConfig) configuration class: [SwitchTransformersForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/switch_transformers#transformers.SwitchTransformersForConditionalGeneration) (SwitchTransformersConfig model) - [T5Config](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Config) configuration class: [T5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5ForConditionalGeneration) (T5Config model) - [T5Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2Config) configuration class: [T5Gemma2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForConditionalGeneration) (T5Gemma2Config model) - [T5GemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaConfig) configuration class: [T5GemmaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaForConditionalGeneration) (T5GemmaConfig model) - [TapasConfig](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasConfig) configuration class: [TapasForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasForMaskedLM) (TapasConfig model) - [UniSpeechConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechConfig) configuration class: [UniSpeechForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechForPreTraining) (UniSpeechConfig model) - [UniSpeechSatConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatConfig) configuration class: [UniSpeechSatForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatForPreTraining) (UniSpeechSatConfig model) - [ViTMAEConfig](/docs/transformers/v5.23.0/ko/model_doc/vit_mae#transformers.ViTMAEConfig) configuration class: [ViTMAEForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/vit_mae#transformers.ViTMAEForPreTraining) (ViTMAEConfig model) - [VibeVoiceAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrConfig) configuration class: [VibeVoiceAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrForConditionalGeneration) (VibeVoiceAsrConfig model) - [VideoLlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaConfig) configuration class: [VideoLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaForConditionalGeneration) (VideoLlavaConfig model) - [VideoMAEConfig](/docs/transformers/v5.23.0/ko/model_doc/videomae#transformers.VideoMAEConfig) configuration class: [VideoMAEForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/videomae#transformers.VideoMAEForPreTraining) (VideoMAEConfig model) - [VipLlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaConfig) configuration class: [VipLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaForConditionalGeneration) (VipLlavaConfig model) - [VisualBertConfig](/docs/transformers/v5.23.0/ko/model_doc/visual_bert#transformers.VisualBertConfig) configuration class: [VisualBertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/visual_bert#transformers.VisualBertForPreTraining) (VisualBertConfig model) - [VoxtralConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralConfig) configuration class: [VoxtralForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralForConditionalGeneration) (VoxtralConfig model) - [VoxtralRealtimeConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeConfig) configuration class: [VoxtralRealtimeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeForConditionalGeneration) (VoxtralRealtimeConfig model) - [Wav2Vec2Config](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Config) configuration class: [Wav2Vec2ForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2ForPreTraining) (Wav2Vec2Config model) - [Wav2Vec2ConformerConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerConfig) configuration class: [Wav2Vec2ConformerForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerForPreTraining) (Wav2Vec2ConformerConfig model) - [XLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMConfig) configuration class: [XLMWithLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMWithLMHeadModel) (XLMConfig model) - [XLMRobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaConfig) configuration class: [XLMRobertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForMaskedLM) (XLMRobertaConfig model) - [XLMRobertaXLConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLConfig) configuration class: [XLMRobertaXLForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForMaskedLM) (XLMRobertaXLConfig model) - [XLNetConfig](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetConfig) configuration class: [XLNetLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetLMHeadModel) (XLNetConfig model) - [XmodConfig](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodConfig) configuration class: [XmodForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForMaskedLM) (XmodConfig model) - [xLSTMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlstm#transformers.xLSTMConfig) configuration class: [xLSTMForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xlstm#transformers.xLSTMForCausalLM) (xLSTMConfig model)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForPreTraining

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForPreTraining.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForPreTraining.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **albert** -- [AlbertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForPreTraining) (AlbertConfig 모델)
- **audioflamingo3** -- [AudioFlamingo3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3ForConditionalGeneration) (AudioFlamingo3Config 모델)
- **bart** -- [BartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForConditionalGeneration) (BartConfig 모델)
- **bert** -- [BertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForPreTraining) (BertConfig 모델)
- **big_bird** -- [BigBirdForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForPreTraining) (BigBirdConfig 모델)
- **bloom** -- [BloomForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomForCausalLM) (BloomConfig 모델)
- **camembert** -- [CamembertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForMaskedLM) (CamembertConfig 모델)
- **colmodernvbert** -- [ColModernVBertForRetrieval](/docs/transformers/v5.23.0/ko/model_doc/colmodernvbert#transformers.ColModernVBertForRetrieval) (ColModernVBertConfig 모델)
- **colpali** -- [ColPaliForRetrieval](/docs/transformers/v5.23.0/ko/model_doc/colpali#transformers.ColPaliForRetrieval) (ColPaliConfig 모델)
- **colqwen2** -- [ColQwen2ForRetrieval](/docs/transformers/v5.23.0/ko/model_doc/colqwen2#transformers.ColQwen2ForRetrieval) (ColQwen2Config 모델)
- **ctrl** -- [CTRLLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLLMHeadModel) (CTRLConfig 모델)
- **data2vec-text** -- [Data2VecTextForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForMaskedLM) (Data2VecTextConfig 모델)
- **deberta** -- [DebertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaForMaskedLM) (DebertaConfig 모델)
- **deberta-v2** -- [DebertaV2ForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForMaskedLM) (DebertaV2Config 모델)
- **distilbert** -- [DistilBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForMaskedLM) (DistilBertConfig 모델)
- **electra** -- [ElectraForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForPreTraining) (ElectraConfig 모델)
- **ernie** -- [ErnieForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForPreTraining) (ErnieConfig 모델)
- **evolla** -- [EvollaForProteinText2Text](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaForProteinText2Text) (EvollaConfig 모델)
- **exaone4** -- [Exaone4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4ForCausalLM) (Exaone4Config 모델)
- **exaone_moe** -- [ExaoneMoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/exaone_moe#transformers.ExaoneMoeForCausalLM) (ExaoneMoeConfig 모델)
- **falcon_mamba** -- [FalconMambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/falcon_mamba#transformers.FalconMambaForCausalLM) (FalconMambaConfig 모델)
- **flaubert** -- [FlaubertWithLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertWithLMHeadModel) (FlaubertConfig 모델)
- **flava** -- [FlavaForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/flava#transformers.FlavaForPreTraining) (FlavaConfig 모델)
- **florence2** -- [Florence2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2ForConditionalGeneration) (Florence2Config 모델)
- **fnet** -- [FNetForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForPreTraining) (FNetConfig 모델)
- **fsmt** -- [FSMTForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fsmt#transformers.FSMTForConditionalGeneration) (FSMTConfig 모델)
- **fun_asr_nano** -- [FunAsrNanoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoForConditionalGeneration) (FunAsrNanoConfig 모델)
- **funnel** -- [FunnelForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForPreTraining) (FunnelConfig 모델)
- **gemma3** -- [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (Gemma3Config 모델)
- **gemma4** -- [Gemma4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4ForConditionalGeneration) (Gemma4Config 모델)
- **gemma4_unified** -- [Gemma4UnifiedForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedForConditionalGeneration) (Gemma4UnifiedConfig 모델)
- **glmasr** -- [GlmAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrForConditionalGeneration) (GlmAsrConfig 모델)
- **gpt-sw3** -- [GPT2LMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2LMHeadModel) (GPT2Config 모델)
- **gpt2** -- [GPT2LMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2LMHeadModel) (GPT2Config 모델)
- **gpt_bigcode** -- [GPTBigCodeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeForCausalLM) (GPTBigCodeConfig 모델)
- **hiera** -- [HieraForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/hiera#transformers.HieraForPreTraining) (HieraConfig 모델)
- **ibert** -- [IBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForMaskedLM) (IBertConfig 모델)
- **idefics** -- [IdeficsForVisionText2Text](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsForVisionText2Text) (IdeficsConfig 모델)
- **idefics2** -- [Idefics2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2ForConditionalGeneration) (Idefics2Config 모델)
- **idefics3** -- [Idefics3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3ForConditionalGeneration) (Idefics3Config 모델)
- **inkling2** -- [Inkling2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2ForConditionalGeneration) (Inkling2Config 모델)
- **inkling_mm_model** -- [InklingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingForConditionalGeneration) (InklingConfig 모델)
- **janus** -- [JanusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusForConditionalGeneration) (JanusConfig 모델)
- **layoutlm** -- [LayoutLMForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMForMaskedLM) (LayoutLMConfig 모델)
- **llava** -- [LlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaForConditionalGeneration) (LlavaConfig 모델)
- **llava_next** -- [LlavaNextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextForConditionalGeneration) (LlavaNextConfig 모델)
- **llava_next_video** -- [LlavaNextVideoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoForConditionalGeneration) (LlavaNextVideoConfig 모델)
- **llava_onevision** -- [LlavaOnevisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionForConditionalGeneration) (LlavaOnevisionConfig 모델)
- **longformer** -- [LongformerForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForMaskedLM) (LongformerConfig 모델)
- **luke** -- [LukeForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForMaskedLM) (LukeConfig 모델)
- **lxmert** -- [LxmertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.LxmertForPreTraining) (LxmertConfig 모델)
- **mamba** -- [MambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mamba#transformers.MambaForCausalLM) (MambaConfig 모델)
- **mamba2** -- [Mamba2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mamba2#transformers.Mamba2ForCausalLM) (Mamba2Config 모델)
- **megatron-bert** -- [MegatronBertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForPreTraining) (MegatronBertConfig 모델)
- **mistral3** -- [Mistral3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3ForConditionalGeneration) (Mistral3Config 모델)
- **mistral4** -- [Mistral4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4ForCausalLM) (Mistral4Config 모델)
- **mistral_large4** -- [MistralLarge4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4ForConditionalGeneration) (MistralLarge4Config 모델)
- **mllama** -- [MllamaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaForConditionalGeneration) (MllamaConfig 모델)
- **mobilebert** -- [MobileBertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForPreTraining) (MobileBertConfig 모델)
- **moss_transcribe_diarize** -- [MossTranscribeDiarizeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeForConditionalGeneration) (MossTranscribeDiarizeConfig 모델)
- **mpnet** -- [MPNetForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForMaskedLM) (MPNetConfig 모델)
- **mpt** -- [MptForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptForCausalLM) (MptConfig 모델)
- **mra** -- [MraForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForMaskedLM) (MraConfig 모델)
- **musicflamingo** -- [MusicFlamingoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/musicflamingo#transformers.MusicFlamingoForConditionalGeneration) (MusicFlamingoConfig 모델)
- **mvp** -- [MvpForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForConditionalGeneration) (MvpConfig 모델)
- **nanochat** -- [NanoChatForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/nanochat#transformers.NanoChatForCausalLM) (NanoChatConfig 모델)
- **nemo_speechlm** -- [Nemotron3_5TranscribeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeForConditionalGeneration) (Nemotron3_5TranscribeConfig 모델)
- **nllb-moe** -- [NllbMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/nllb-moe#transformers.NllbMoeForConditionalGeneration) (NllbMoeConfig 모델)
- **openai-gpt** -- [OpenAIGPTLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTLMHeadModel) (OpenAIGPTConfig 모델)
- **paligemma** -- [PaliGemmaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaForConditionalGeneration) (PaliGemmaConfig 모델)
- **qwen2_audio** -- [Qwen2AudioForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioForConditionalGeneration) (Qwen2AudioConfig 모델)
- **qwen3_asr** -- [Qwen3ASRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRForConditionalGeneration) (Qwen3ASRConfig 모델)
- **roberta** -- [RobertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForMaskedLM) (RobertaConfig 모델)
- **roberta-prelayernorm** -- [RobertaPreLayerNormForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForMaskedLM) (RobertaPreLayerNormConfig 모델)
- **roc_bert** -- [RoCBertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForPreTraining) (RoCBertConfig 모델)
- **rwkv** -- [RwkvForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/rwkv#transformers.RwkvForCausalLM) (RwkvConfig 모델)
- **splinter** -- [SplinterForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/splinter#transformers.SplinterForPreTraining) (SplinterConfig 모델)
- **squeezebert** -- [SqueezeBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForMaskedLM) (SqueezeBertConfig 모델)
- **switch_transformers** -- [SwitchTransformersForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/switch_transformers#transformers.SwitchTransformersForConditionalGeneration) (SwitchTransformersConfig 모델)
- **t5** -- [T5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5ForConditionalGeneration) (T5Config 모델)
- **t5gemma** -- [T5GemmaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaForConditionalGeneration) (T5GemmaConfig 모델)
- **t5gemma2** -- [T5Gemma2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForConditionalGeneration) (T5Gemma2Config 모델)
- **tapas** -- [TapasForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasForMaskedLM) (TapasConfig 모델)
- **unispeech** -- [UniSpeechForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechForPreTraining) (UniSpeechConfig 모델)
- **unispeech-sat** -- [UniSpeechSatForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatForPreTraining) (UniSpeechSatConfig 모델)
- **vibevoice_asr** -- [VibeVoiceAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrForConditionalGeneration) (VibeVoiceAsrConfig 모델)
- **video_llava** -- [VideoLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaForConditionalGeneration) (VideoLlavaConfig 모델)
- **videomae** -- [VideoMAEForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/videomae#transformers.VideoMAEForPreTraining) (VideoMAEConfig 모델)
- **vipllava** -- [VipLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaForConditionalGeneration) (VipLlavaConfig 모델)
- **visual_bert** -- [VisualBertForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/visual_bert#transformers.VisualBertForPreTraining) (VisualBertConfig 모델)
- **vit_mae** -- [ViTMAEForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/vit_mae#transformers.ViTMAEForPreTraining) (ViTMAEConfig 모델)
- **voxtral** -- [VoxtralForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralForConditionalGeneration) (VoxtralConfig 모델)
- **voxtral_realtime** -- [VoxtralRealtimeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeForConditionalGeneration) (VoxtralRealtimeConfig 모델)
- **wav2vec2** -- [Wav2Vec2ForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2ForPreTraining) (Wav2Vec2Config 모델)
- **wav2vec2-conformer** -- [Wav2Vec2ConformerForPreTraining](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerForPreTraining) (Wav2Vec2ConformerConfig 모델)
- **xlm** -- [XLMWithLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMWithLMHeadModel) (XLMConfig 모델)
- **xlm-roberta** -- [XLMRobertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForMaskedLM) (XLMRobertaConfig 모델)
- **xlm-roberta-xl** -- [XLMRobertaXLForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForMaskedLM) (XLMRobertaXLConfig 모델)
- **xlnet** -- [XLNetLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetLMHeadModel) (XLNetConfig 모델)
- **xlstm** -- [xLSTMForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xlstm#transformers.xLSTMForCausalLM) (xLSTMConfig 모델)
- **xmod** -- [XmodForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForMaskedLM) (XmodConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForPreTraining

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForPreTraining.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForPreTraining.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

## 자연어 처리

다음은 자연어 처리 작업을 위해 사용할 수 있는 자동 클래스입니다.

### AutoModelForCausalLM[[transformers.AutoModelForCausalLM]]

#### transformers.AutoModelForCausalLM[[transformers.AutoModelForCausalLM]]

```python
transformers.AutoModelForCausalLM(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2347)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 범용 모델 클래스입니다(인과적 언어 모델링 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForCausalLM.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : The model class to instantiate is selected based on the configuration class:  - [AXK1Config](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1Config) configuration class: [AXK1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1ForCausalLM) (AXK1Config model) - [AXK2Config](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2Config) configuration class: [AXK2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2ForCausalLM) (AXK2Config model) - [AfmoeConfig](/docs/transformers/v5.23.0/ko/model_doc/afmoe#transformers.AfmoeConfig) configuration class: [AfmoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/afmoe#transformers.AfmoeForCausalLM) (AfmoeConfig model) - [ApertusConfig](/docs/transformers/v5.23.0/ko/model_doc/apertus#transformers.ApertusConfig) configuration class: [ApertusForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/apertus#transformers.ApertusForCausalLM) (ApertusConfig model) - [ArceeConfig](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeConfig) configuration class: [ArceeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeForCausalLM) (ArceeConfig model) - [AriaTextConfig](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaTextConfig) configuration class: [AriaTextForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaTextForCausalLM) (AriaTextConfig model) - [BailingHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/bailing_hybrid#transformers.BailingHybridConfig) configuration class: [BailingHybridForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bailing_hybrid#transformers.BailingHybridForCausalLM) (BailingHybridConfig model) - [BambaConfig](/docs/transformers/v5.23.0/ko/model_doc/bamba#transformers.BambaConfig) configuration class: [BambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bamba#transformers.BambaForCausalLM) (BambaConfig model) - [BartConfig](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartConfig) configuration class: [BartForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForCausalLM) (BartConfig model) - [BeamConfig](/docs/transformers/v5.23.0/ko/model_doc/beam#transformers.BeamConfig) configuration class: [BeamForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/beam#transformers.BeamForCausalLM) (BeamConfig model) - [BertConfig](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertConfig) configuration class: [BertLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertLMHeadModel) (BertConfig model) - [BertGenerationConfig](/docs/transformers/v5.23.0/ko/model_doc/bert-generation#transformers.BertGenerationConfig) configuration class: [BertGenerationDecoder](/docs/transformers/v5.23.0/ko/model_doc/bert-generation#transformers.BertGenerationDecoder) (BertGenerationConfig model) - [BigBirdConfig](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdConfig) configuration class: [BigBirdForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForCausalLM) (BigBirdConfig model) - [BigBirdPegasusConfig](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusConfig) configuration class: [BigBirdPegasusForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusForCausalLM) (BigBirdPegasusConfig model) - [BioGptConfig](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptConfig) configuration class: [BioGptForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptForCausalLM) (BioGptConfig model) - [BitNetConfig](/docs/transformers/v5.23.0/ko/model_doc/bitnet#transformers.BitNetConfig) configuration class: [BitNetForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bitnet#transformers.BitNetForCausalLM) (BitNetConfig model) - [BlenderbotConfig](/docs/transformers/v5.23.0/ko/model_doc/blenderbot#transformers.BlenderbotConfig) configuration class: [BlenderbotForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/blenderbot#transformers.BlenderbotForCausalLM) (BlenderbotConfig model) - [BlenderbotSmallConfig](/docs/transformers/v5.23.0/ko/model_doc/blenderbot-small#transformers.BlenderbotSmallConfig) configuration class: [BlenderbotSmallForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/blenderbot-small#transformers.BlenderbotSmallForCausalLM) (BlenderbotSmallConfig model) - [BloomConfig](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomConfig) configuration class: [BloomForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomForCausalLM) (BloomConfig model) - [BltConfig](/docs/transformers/v5.23.0/ko/model_doc/blt#transformers.BltConfig) configuration class: [BltForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/blt#transformers.BltForCausalLM) (BltConfig model) - [CTRLConfig](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLConfig) configuration class: [CTRLLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLLMHeadModel) (CTRLConfig model) - [CamembertConfig](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertConfig) configuration class: [CamembertForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForCausalLM) (CamembertConfig model) - [CodeGenConfig](/docs/transformers/v5.23.0/ko/model_doc/codegen#transformers.CodeGenConfig) configuration class: [CodeGenForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/codegen#transformers.CodeGenForCausalLM) (CodeGenConfig model) - [Cohere2Config](/docs/transformers/v5.23.0/ko/model_doc/cohere2#transformers.Cohere2Config) configuration class: [Cohere2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cohere2#transformers.Cohere2ForCausalLM) (Cohere2Config model) - [Cohere2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere2_moe#transformers.Cohere2MoeConfig) configuration class: [Cohere2MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cohere2_moe#transformers.Cohere2MoeForCausalLM) (Cohere2MoeConfig model) - [CohereCompassTextConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassTextConfig) configuration class: [CohereCompassForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassForCausalLM) (CohereCompassTextConfig model) - [CohereConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere#transformers.CohereConfig) configuration class: [CohereForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cohere#transformers.CohereForCausalLM) (CohereConfig model) - [CpmAntConfig](/docs/transformers/v5.23.0/ko/model_doc/cpmant#transformers.CpmAntConfig) configuration class: [CpmAntForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cpmant#transformers.CpmAntForCausalLM) (CpmAntConfig model) - [CwmConfig](/docs/transformers/v5.23.0/ko/model_doc/cwm#transformers.CwmConfig) configuration class: [CwmForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cwm#transformers.CwmForCausalLM) (CwmConfig model) - [Data2VecTextConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextConfig) configuration class: [Data2VecTextForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForCausalLM) (Data2VecTextConfig model) - [DbrxConfig](/docs/transformers/v5.23.0/ko/model_doc/dbrx#transformers.DbrxConfig) configuration class: [DbrxForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/dbrx#transformers.DbrxForCausalLM) (DbrxConfig model) - [DeepseekV2Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v2#transformers.DeepseekV2Config) configuration class: [DeepseekV2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v2#transformers.DeepseekV2ForCausalLM) (DeepseekV2Config model) - [DeepseekV32Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v32#transformers.DeepseekV32Config) configuration class: [DeepseekV32ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v32#transformers.DeepseekV32ForCausalLM) (DeepseekV32Config model) - [DeepseekV3Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3Config) configuration class: [DeepseekV3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3ForCausalLM) (DeepseekV3Config model) - [DeepseekV4Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v4#transformers.DeepseekV4Config) configuration class: [DeepseekV4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v4#transformers.DeepseekV4ForCausalLM) (DeepseekV4Config model) - [DiffLlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaConfig) configuration class: [DiffLlamaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaForCausalLM) (DiffLlamaConfig model) - [DogeConfig](/docs/transformers/v5.23.0/ko/model_doc/doge#transformers.DogeConfig) configuration class: [DogeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/doge#transformers.DogeForCausalLM) (DogeConfig model) - [Dots1Config](/docs/transformers/v5.23.0/ko/model_doc/dots1#transformers.Dots1Config) configuration class: [Dots1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/dots1#transformers.Dots1ForCausalLM) (Dots1Config model) - [Dots3NoteConfig](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteConfig) configuration class: [Dots3NoteForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteForCausalLM) (Dots3NoteConfig model) - [Dots3NoteTextConfig](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteTextConfig) configuration class: [Dots3NoteForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteForCausalLM) (Dots3NoteTextConfig model) - [ElectraConfig](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraConfig) configuration class: [ElectraForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForCausalLM) (ElectraConfig model) - [Emu3Config](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3Config) configuration class: [Emu3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3ForCausalLM) (Emu3Config model) - [Ernie4_5Config](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5#transformers.Ernie4_5Config) configuration class: [Ernie4_5ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5#transformers.Ernie4_5ForCausalLM) (Ernie4_5Config model) - [Ernie4_5_MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_moe#transformers.Ernie4_5_MoeConfig) configuration class: [Ernie4_5_MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_moe#transformers.Ernie4_5_MoeForCausalLM) (Ernie4_5_MoeConfig model) - [ErnieConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieConfig) configuration class: [ErnieForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForCausalLM) (ErnieConfig model) - [Exaone4Config](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4Config) configuration class: [Exaone4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4ForCausalLM) (Exaone4Config model) - [ExaoneMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/exaone_moe#transformers.ExaoneMoeConfig) configuration class: [ExaoneMoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/exaone_moe#transformers.ExaoneMoeForCausalLM) (ExaoneMoeConfig model) - [FalconConfig](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconConfig) configuration class: [FalconForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconForCausalLM) (FalconConfig model) - [FalconH1Config](/docs/transformers/v5.23.0/ko/model_doc/falcon_h1#transformers.FalconH1Config) configuration class: [FalconH1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/falcon_h1#transformers.FalconH1ForCausalLM) (FalconH1Config model) - [FalconMambaConfig](/docs/transformers/v5.23.0/ko/model_doc/falcon_mamba#transformers.FalconMambaConfig) configuration class: [FalconMambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/falcon_mamba#transformers.FalconMambaForCausalLM) (FalconMambaConfig model) - [FlexOlmoConfig](/docs/transformers/v5.23.0/ko/model_doc/flex_olmo#transformers.FlexOlmoConfig) configuration class: [FlexOlmoForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/flex_olmo#transformers.FlexOlmoForCausalLM) (FlexOlmoConfig model) - [FlintConfig](/docs/transformers/v5.23.0/ko/model_doc/flint#transformers.FlintConfig) configuration class: [FlintForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/flint#transformers.FlintForCausalLM) (FlintConfig model) - [FuyuConfig](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuConfig) configuration class: [FuyuForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuForCausalLM) (FuyuConfig model) - [GPT2Config](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Config) configuration class: [GPT2LMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2LMHeadModel) (GPT2Config model) - [GPTBigCodeConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeConfig) configuration class: [GPTBigCodeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeForCausalLM) (GPTBigCodeConfig model) - [GPTJConfig](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJConfig) configuration class: [GPTJForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJForCausalLM) (GPTJConfig model) - [GPTNeoConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoConfig) configuration class: [GPTNeoForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoForCausalLM) (GPTNeoConfig model) - [GPTNeoXConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXConfig) configuration class: [GPTNeoXForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXForCausalLM) (GPTNeoXConfig model) - [GPTNeoXJapaneseConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox_japanese#transformers.GPTNeoXJapaneseConfig) configuration class: [GPTNeoXJapaneseForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox_japanese#transformers.GPTNeoXJapaneseForCausalLM) (GPTNeoXJapaneseConfig model) - [Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2Config) configuration class: [Gemma2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2ForCausalLM) (Gemma2Config model) - [Gemma3Config](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Config) configuration class: [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (Gemma3Config model) - [Gemma3TextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3TextConfig) configuration class: [Gemma3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForCausalLM) (Gemma3TextConfig model) - [Gemma3nConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nConfig) configuration class: [Gemma3nForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nForConditionalGeneration) (Gemma3nConfig model) - [Gemma3nTextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nTextConfig) configuration class: [Gemma3nForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nForCausalLM) (Gemma3nTextConfig model) - [Gemma4AssistantConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_assistant#transformers.Gemma4AssistantConfig) configuration class: [Gemma4AssistantForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma4_assistant#transformers.Gemma4AssistantForCausalLM) (Gemma4AssistantConfig model) - [Gemma4Config](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4Config) configuration class: [Gemma4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4ForConditionalGeneration) (Gemma4Config model) - [Gemma4TextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4TextConfig) configuration class: [Gemma4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4ForCausalLM) (Gemma4TextConfig model) - [Gemma4UnifiedAssistantConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified_assistant#transformers.Gemma4UnifiedAssistantConfig) configuration class: [Gemma4UnifiedAssistantForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified_assistant#transformers.Gemma4UnifiedAssistantForCausalLM) (Gemma4UnifiedAssistantConfig model) - [Gemma4UnifiedConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedConfig) configuration class: [Gemma4UnifiedForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedForConditionalGeneration) (Gemma4UnifiedConfig model) - [Gemma4UnifiedTextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedTextConfig) configuration class: [Gemma4UnifiedForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedForCausalLM) (Gemma4UnifiedTextConfig model) - [GemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaConfig) configuration class: [GemmaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaForCausalLM) (GemmaConfig model) - [GitConfig](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitConfig) configuration class: [GitForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitForCausalLM) (GitConfig model) - [Glm4Config](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4Config) configuration class: [Glm4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4ForCausalLM) (Glm4Config model) - [Glm4MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe#transformers.Glm4MoeConfig) configuration class: [Glm4MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe#transformers.Glm4MoeForCausalLM) (Glm4MoeConfig model) - [Glm4MoeLiteConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe_lite#transformers.Glm4MoeLiteConfig) configuration class: [Glm4MoeLiteForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe_lite#transformers.Glm4MoeLiteForCausalLM) (Glm4MoeLiteConfig model) - [GlmConfig](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmConfig) configuration class: [GlmForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmForCausalLM) (GlmConfig model) - [GlmMoeDsaConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_moe_dsa#transformers.GlmMoeDsaConfig) configuration class: [GlmMoeDsaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/glm_moe_dsa#transformers.GlmMoeDsaForCausalLM) (GlmMoeDsaConfig model) - [GotOcr2Config](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2Config) configuration class: [GotOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2ForConditionalGeneration) (GotOcr2Config model) - [GptOssConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssConfig) configuration class: [GptOssForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssForCausalLM) (GptOssConfig model) - [GraniteConfig](/docs/transformers/v5.23.0/ko/model_doc/granite#transformers.GraniteConfig) configuration class: [GraniteForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granite#transformers.GraniteForCausalLM) (GraniteConfig model) - [GraniteMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoe#transformers.GraniteMoeConfig) configuration class: [GraniteMoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granitemoe#transformers.GraniteMoeForCausalLM) (GraniteMoeConfig model) - [GraniteMoeHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoehybrid#transformers.GraniteMoeHybridConfig) configuration class: [GraniteMoeHybridForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granitemoehybrid#transformers.GraniteMoeHybridForCausalLM) (GraniteMoeHybridConfig model) - [GraniteMoeSWAConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoe_swa#transformers.GraniteMoeSWAConfig) configuration class: [GraniteMoeSWAForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granitemoe_swa#transformers.GraniteMoeSWAForCausalLM) (GraniteMoeSWAConfig model) - [GraniteMoeSharedConfig](/docs/transformers/v5.23.0/ko/model_doc/granitemoeshared#transformers.GraniteMoeSharedConfig) configuration class: [GraniteMoeSharedForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granitemoeshared#transformers.GraniteMoeSharedForCausalLM) (GraniteMoeSharedConfig model) - [GraniteSWAConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_swa#transformers.GraniteSWAConfig) configuration class: [GraniteSWAForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granite_swa#transformers.GraniteSWAForCausalLM) (GraniteSWAConfig model) - [HYV3Config](/docs/transformers/v5.23.0/ko/model_doc/hy_v3#transformers.HYV3Config) configuration class: [HYV3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hy_v3#transformers.HYV3ForCausalLM) (HYV3Config model) - [HYV4Config](/docs/transformers/v5.23.0/ko/model_doc/hy_v4#transformers.HYV4Config) configuration class: [HYV4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hy_v4#transformers.HYV4ForCausalLM) (HYV4Config model) - [HeliumConfig](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumConfig) configuration class: [HeliumForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumForCausalLM) (HeliumConfig model) - [HrmTextConfig](/docs/transformers/v5.23.0/ko/model_doc/hrm_text#transformers.HrmTextConfig) configuration class: [HrmTextForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hrm_text#transformers.HrmTextForCausalLM) (HrmTextConfig model) - [HunYuanDenseV1Config](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_dense#transformers.HunYuanDenseV1Config) configuration class: [HunYuanDenseV1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_dense#transformers.HunYuanDenseV1ForCausalLM) (HunYuanDenseV1Config model) - [HunYuanMoEV1Config](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_moe#transformers.HunYuanMoEV1Config) configuration class: [HunYuanMoEV1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_moe#transformers.HunYuanMoEV1ForCausalLM) (HunYuanMoEV1Config model) - [HyperCLOVAXConfig](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax#transformers.HyperCLOVAXConfig) configuration class: [HyperCLOVAXForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax#transformers.HyperCLOVAXForCausalLM) (HyperCLOVAXConfig model) - [IQuestCoderConfig](/docs/transformers/v5.23.0/ko/model_doc/iquestcoder#transformers.IQuestCoderConfig) configuration class: [LlamaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaForCausalLM) (IQuestCoderConfig model) - [Inkling2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2TextConfig) configuration class: [Inkling2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2ForCausalLM) (Inkling2TextConfig model) - [InklingTextConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingTextConfig) configuration class: [InklingForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingForCausalLM) (InklingTextConfig model) - [InternLM3Config](/docs/transformers/v5.23.0/ko/model_doc/internlm3#transformers.InternLM3Config) configuration class: [LlamaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaForCausalLM) (InternLM3Config model) - [Jais2Config](/docs/transformers/v5.23.0/ko/model_doc/jais2#transformers.Jais2Config) configuration class: [Jais2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/jais2#transformers.Jais2ForCausalLM) (Jais2Config model) - [JambaConfig](/docs/transformers/v5.23.0/ko/model_doc/jamba#transformers.JambaConfig) configuration class: [JambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/jamba#transformers.JambaForCausalLM) (JambaConfig model) - [JetMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/jetmoe#transformers.JetMoeConfig) configuration class: [JetMoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/jetmoe#transformers.JetMoeForCausalLM) (JetMoeConfig model) - [KimiLinearConfig](/docs/transformers/v5.23.0/ko/model_doc/kimi_linear#transformers.KimiLinearConfig) configuration class: [KimiLinearForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/kimi_linear#transformers.KimiLinearForCausalLM) (KimiLinearConfig model) - [Kolibri1Config](/docs/transformers/v5.23.0/ko/model_doc/kolibri1#transformers.Kolibri1Config) configuration class: [Kolibri1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/kolibri1#transformers.Kolibri1ForCausalLM) (Kolibri1Config model) - [LagunaConfig](/docs/transformers/v5.23.0/ko/model_doc/laguna#transformers.LagunaConfig) configuration class: [LagunaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/laguna#transformers.LagunaForCausalLM) (LagunaConfig model) - [Lfm2Config](/docs/transformers/v5.23.0/ko/model_doc/lfm2#transformers.Lfm2Config) configuration class: [Lfm2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/lfm2#transformers.Lfm2ForCausalLM) (Lfm2Config model) - [Lfm2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm2_moe#transformers.Lfm2MoeConfig) configuration class: [Lfm2MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/lfm2_moe#transformers.Lfm2MoeForCausalLM) (Lfm2MoeConfig model) - [Lfm3Config](/docs/transformers/v5.23.0/ko/model_doc/lfm3#transformers.Lfm3Config) configuration class: [Lfm3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/lfm3#transformers.Lfm3ForCausalLM) (Lfm3Config model) - [Llama4Config](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4Config) configuration class: [Llama4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4ForCausalLM) (Llama4Config model) - [Llama4TextConfig](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4TextConfig) configuration class: [Llama4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4ForCausalLM) (Llama4TextConfig model) - [LlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaConfig) configuration class: [LlamaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaForCausalLM) (LlamaConfig model) - [LongcatFlashConfig](/docs/transformers/v5.23.0/ko/model_doc/longcat_flash#transformers.LongcatFlashConfig) configuration class: [LongcatFlashForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/longcat_flash#transformers.LongcatFlashForCausalLM) (LongcatFlashConfig model) - [MBartConfig](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartConfig) configuration class: [MBartForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartForCausalLM) (MBartConfig model) - [Mamba2Config](/docs/transformers/v5.23.0/ko/model_doc/mamba2#transformers.Mamba2Config) configuration class: [Mamba2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mamba2#transformers.Mamba2ForCausalLM) (Mamba2Config model) - [MambaConfig](/docs/transformers/v5.23.0/ko/model_doc/mamba#transformers.MambaConfig) configuration class: [MambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mamba#transformers.MambaForCausalLM) (MambaConfig model) - [MarianConfig](/docs/transformers/v5.23.0/ko/model_doc/marian#transformers.MarianConfig) configuration class: [MarianForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/marian#transformers.MarianForCausalLM) (MarianConfig model) - [MegatronBertConfig](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertConfig) configuration class: [MegatronBertForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForCausalLM) (MegatronBertConfig model) - [MellumConfig](/docs/transformers/v5.23.0/ko/model_doc/mellum#transformers.MellumConfig) configuration class: [MellumForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mellum#transformers.MellumForCausalLM) (MellumConfig model) - [MiMoV2FlashConfig](/docs/transformers/v5.23.0/ko/model_doc/mimo_v2_flash#transformers.MiMoV2FlashConfig) configuration class: [MiMoV2FlashForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mimo_v2_flash#transformers.MiMoV2FlashForCausalLM) (MiMoV2FlashConfig model) - [MiniCPM3Config](/docs/transformers/v5.23.0/ko/model_doc/minicpm3#transformers.MiniCPM3Config) configuration class: [MiniCPM3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/minicpm3#transformers.MiniCPM3ForCausalLM) (MiniCPM3Config model) - [MiniMaxConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxConfig) configuration class: [MiniMaxForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxForCausalLM) (MiniMaxConfig model) - [MiniMaxM2Config](/docs/transformers/v5.23.0/ko/model_doc/minimax_m2#transformers.MiniMaxM2Config) configuration class: [MiniMaxM2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/minimax_m2#transformers.MiniMaxM2ForCausalLM) (MiniMaxM2Config model) - [MiniMaxM3VLTextConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLTextConfig) configuration class: [MiniMaxM3VLForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLForCausalLM) (MiniMaxM3VLTextConfig model) - [Ministral3Config](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3Config) configuration class: [Ministral3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3ForCausalLM) (Ministral3Config model) - [MinistralConfig](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralConfig) configuration class: [MinistralForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralForCausalLM) (MinistralConfig model) - [MistralConfig](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralConfig) configuration class: [MistralForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralForCausalLM) (MistralConfig model) - [MixtralConfig](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralConfig) configuration class: [MixtralForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralForCausalLM) (MixtralConfig model) - [MllamaConfig](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaConfig) configuration class: [MllamaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaForCausalLM) (MllamaConfig model) - [ModernBertDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/modernbert-decoder#transformers.ModernBertDecoderConfig) configuration class: [ModernBertDecoderForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/modernbert-decoder#transformers.ModernBertDecoderForCausalLM) (ModernBertDecoderConfig model) - [MoshiConfig](/docs/transformers/v5.23.0/ko/model_doc/moshi#transformers.MoshiConfig) configuration class: [MoshiForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/moshi#transformers.MoshiForCausalLM) (MoshiConfig model) - [MossTranscribeDiarizeConfig](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeConfig) configuration class: [MossTranscribeDiarizeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeForConditionalGeneration) (MossTranscribeDiarizeConfig model) - [MptConfig](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptConfig) configuration class: [MptForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptForCausalLM) (MptConfig model) - [MusicgenConfig](/docs/transformers/v5.23.0/ko/model_doc/musicgen#transformers.MusicgenConfig) configuration class: [MusicgenForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/musicgen#transformers.MusicgenForCausalLM) (MusicgenConfig model) - [MusicgenMelodyConfig](/docs/transformers/v5.23.0/ko/model_doc/musicgen_melody#transformers.MusicgenMelodyConfig) configuration class: [MusicgenMelodyForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/musicgen_melody#transformers.MusicgenMelodyForCausalLM) (MusicgenMelodyConfig model) - [MvpConfig](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpConfig) configuration class: [MvpForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForCausalLM) (MvpConfig model) - [NanoChatConfig](/docs/transformers/v5.23.0/ko/model_doc/nanochat#transformers.NanoChatConfig) configuration class: [NanoChatForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/nanochat#transformers.NanoChatForCausalLM) (NanoChatConfig model) - [NemotronConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronConfig) configuration class: [NemotronForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronForCausalLM) (NemotronConfig model) - [NemotronHConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h#transformers.NemotronHConfig) configuration class: [NemotronHForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h#transformers.NemotronHForCausalLM) (NemotronHConfig model) - [OPTConfig](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTConfig) configuration class: [OPTForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTForCausalLM) (OPTConfig model) - [Olmo2Config](/docs/transformers/v5.23.0/ko/model_doc/olmo2#transformers.Olmo2Config) configuration class: [Olmo2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/olmo2#transformers.Olmo2ForCausalLM) (Olmo2Config model) - [Olmo3Config](/docs/transformers/v5.23.0/ko/model_doc/olmo3#transformers.Olmo3Config) configuration class: [Olmo3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/olmo3#transformers.Olmo3ForCausalLM) (Olmo3Config model) - [OlmoConfig](/docs/transformers/v5.23.0/ko/model_doc/olmo#transformers.OlmoConfig) configuration class: [OlmoForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/olmo#transformers.OlmoForCausalLM) (OlmoConfig model) - [OlmoHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/olmo_hybrid#transformers.OlmoHybridConfig) configuration class: [OlmoHybridForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/olmo_hybrid#transformers.OlmoHybridForCausalLM) (OlmoHybridConfig model) - [OlmoeConfig](/docs/transformers/v5.23.0/ko/model_doc/olmoe#transformers.OlmoeConfig) configuration class: [OlmoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/olmoe#transformers.OlmoeForCausalLM) (OlmoeConfig model) - [OpenAIGPTConfig](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTConfig) configuration class: [OpenAIGPTLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTLMHeadModel) (OpenAIGPTConfig model) - [PLBartConfig](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartConfig) configuration class: [PLBartForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartForCausalLM) (PLBartConfig model) - [PegasusConfig](/docs/transformers/v5.23.0/ko/model_doc/pegasus#transformers.PegasusConfig) configuration class: [PegasusForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/pegasus#transformers.PegasusForCausalLM) (PegasusConfig model) - [PersimmonConfig](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonConfig) configuration class: [PersimmonForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonForCausalLM) (PersimmonConfig model) - [Phi3Config](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3Config) configuration class: [Phi3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3ForCausalLM) (Phi3Config model) - [Phi4MultimodalConfig](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalConfig) configuration class: [Phi4MultimodalForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalForCausalLM) (Phi4MultimodalConfig model) - [PhiConfig](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiConfig) configuration class: [PhiForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiForCausalLM) (PhiConfig model) - [PhimoeConfig](/docs/transformers/v5.23.0/ko/model_doc/phimoe#transformers.PhimoeConfig) configuration class: [PhimoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/phimoe#transformers.PhimoeForCausalLM) (PhimoeConfig model) - [ProphetNetConfig](/docs/transformers/v5.23.0/ko/model_doc/prophetnet#transformers.ProphetNetConfig) configuration class: [ProphetNetForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/prophetnet#transformers.ProphetNetForCausalLM) (ProphetNetConfig model) - [Qwen2Config](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Config) configuration class: [Qwen2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2ForCausalLM) (Qwen2Config model) - [Qwen2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeConfig) configuration class: [Qwen2MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeForCausalLM) (Qwen2MoeConfig model) - [Qwen3Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3Config) configuration class: [Qwen3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3ForCausalLM) (Qwen3Config model) - [Qwen3MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeConfig) configuration class: [Qwen3MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeForCausalLM) (Qwen3MoeConfig model) - [Qwen3NextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextConfig) configuration class: [Qwen3NextForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextForCausalLM) (Qwen3NextConfig model) - [Qwen3_5Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Config) configuration class: [Qwen3_5ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForCausalLM) (Qwen3_5Config model) - [Qwen3_5MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeConfig) configuration class: [Qwen3_5MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeForCausalLM) (Qwen3_5MoeConfig model) - [Qwen3_5MoeTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeTextConfig) configuration class: [Qwen3_5MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeForCausalLM) (Qwen3_5MoeTextConfig model) - [Qwen3_5TextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5TextConfig) configuration class: [Qwen3_5ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForCausalLM) (Qwen3_5TextConfig model) - [Qwen4ExpConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpConfig) configuration class: [Qwen4ExpForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpForCausalLM) (Qwen4ExpConfig model) - [Qwen4ExpTextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpTextConfig) configuration class: [Qwen4ExpForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpForCausalLM) (Qwen4ExpTextConfig model) - [RecurrentGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/recurrent_gemma#transformers.RecurrentGemmaConfig) configuration class: [RecurrentGemmaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/recurrent_gemma#transformers.RecurrentGemmaForCausalLM) (RecurrentGemmaConfig model) - [ReformerConfig](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerConfig) configuration class: [ReformerModelWithLMHead](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerModelWithLMHead) (ReformerConfig model) - [RemBertConfig](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertConfig) configuration class: [RemBertForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForCausalLM) (RemBertConfig model) - [RoCBertConfig](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertConfig) configuration class: [RoCBertForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForCausalLM) (RoCBertConfig model) - [RoFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerConfig) configuration class: [RoFormerForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForCausalLM) (RoFormerConfig model) - [RobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaConfig) configuration class: [RobertaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForCausalLM) (RobertaConfig model) - [RobertaPreLayerNormConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormConfig) configuration class: [RobertaPreLayerNormForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForCausalLM) (RobertaPreLayerNormConfig model) - [RwkvConfig](/docs/transformers/v5.23.0/ko/model_doc/rwkv#transformers.RwkvConfig) configuration class: [RwkvForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/rwkv#transformers.RwkvForCausalLM) (RwkvConfig model) - [SeedOssConfig](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssConfig) configuration class: [SeedOssForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssForCausalLM) (SeedOssConfig model) - [SmolLM3Config](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3Config) configuration class: [SmolLM3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3ForCausalLM) (SmolLM3Config model) - [SolarOpenConfig](/docs/transformers/v5.23.0/ko/model_doc/solar_open#transformers.SolarOpenConfig) configuration class: [SolarOpenForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/solar_open#transformers.SolarOpenForCausalLM) (SolarOpenConfig model) - [StableLmConfig](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmConfig) configuration class: [StableLmForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmForCausalLM) (StableLmConfig model) - [Starcoder2Config](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2Config) configuration class: [Starcoder2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2ForCausalLM) (Starcoder2Config model) - [TrOCRConfig](/docs/transformers/v5.23.0/ko/model_doc/trocr#transformers.TrOCRConfig) configuration class: [TrOCRForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/trocr#transformers.TrOCRForCausalLM) (TrOCRConfig model) - [VaultGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/vaultgemma#transformers.VaultGemmaConfig) configuration class: [VaultGemmaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/vaultgemma#transformers.VaultGemmaForCausalLM) (VaultGemmaConfig model) - [WhisperConfig](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperConfig) configuration class: [WhisperForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperForCausalLM) (WhisperConfig model) - [XGLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xglm#transformers.XGLMConfig) configuration class: [XGLMForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xglm#transformers.XGLMForCausalLM) (XGLMConfig model) - [XLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMConfig) configuration class: [XLMWithLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMWithLMHeadModel) (XLMConfig model) - [XLMRobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaConfig) configuration class: [XLMRobertaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForCausalLM) (XLMRobertaConfig model) - [XLMRobertaXLConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLConfig) configuration class: [XLMRobertaXLForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForCausalLM) (XLMRobertaXLConfig model) - [XLNetConfig](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetConfig) configuration class: [XLNetLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetLMHeadModel) (XLNetConfig model) - [XmodConfig](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodConfig) configuration class: [XmodForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForCausalLM) (XmodConfig model) - [YoutuConfig](/docs/transformers/v5.23.0/ko/model_doc/youtu#transformers.YoutuConfig) configuration class: [YoutuForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/youtu#transformers.YoutuForCausalLM) (YoutuConfig model) - [Zamba2Config](/docs/transformers/v5.23.0/ko/model_doc/zamba2#transformers.Zamba2Config) configuration class: [Zamba2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/zamba2#transformers.Zamba2ForCausalLM) (Zamba2Config model) - [ZambaConfig](/docs/transformers/v5.23.0/ko/model_doc/zamba#transformers.ZambaConfig) configuration class: [ZambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/zamba#transformers.ZambaForCausalLM) (ZambaConfig model) - [ZayaConfig](/docs/transformers/v5.23.0/ko/model_doc/zaya#transformers.ZayaConfig) configuration class: [ZayaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/zaya#transformers.ZayaForCausalLM) (ZayaConfig model) - [xLSTMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlstm#transformers.xLSTMConfig) configuration class: [xLSTMForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xlstm#transformers.xLSTMForCausalLM) (xLSTMConfig model)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForCausalLM

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForCausalLM.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForCausalLM.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **afmoe** -- [AfmoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/afmoe#transformers.AfmoeForCausalLM) (AfmoeConfig 모델)
- **apertus** -- [ApertusForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/apertus#transformers.ApertusForCausalLM) (ApertusConfig 모델)
- **arcee** -- [ArceeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeForCausalLM) (ArceeConfig 모델)
- **aria_text** -- [AriaTextForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaTextForCausalLM) (AriaTextConfig 모델)
- **axk1** -- [AXK1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1ForCausalLM) (AXK1Config 모델)
- **axk2** -- [AXK2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2ForCausalLM) (AXK2Config 모델)
- **bailing_hybrid** -- [BailingHybridForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bailing_hybrid#transformers.BailingHybridForCausalLM) (BailingHybridConfig 모델)
- **bamba** -- [BambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bamba#transformers.BambaForCausalLM) (BambaConfig 모델)
- **bart** -- [BartForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForCausalLM) (BartConfig 모델)
- **beam** -- [BeamForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/beam#transformers.BeamForCausalLM) (BeamConfig 모델)
- **bert** -- [BertLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertLMHeadModel) (BertConfig 모델)
- **bert-generation** -- [BertGenerationDecoder](/docs/transformers/v5.23.0/ko/model_doc/bert-generation#transformers.BertGenerationDecoder) (BertGenerationConfig 모델)
- **big_bird** -- [BigBirdForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForCausalLM) (BigBirdConfig 모델)
- **bigbird_pegasus** -- [BigBirdPegasusForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusForCausalLM) (BigBirdPegasusConfig 모델)
- **biogpt** -- [BioGptForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptForCausalLM) (BioGptConfig 모델)
- **bitnet** -- [BitNetForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bitnet#transformers.BitNetForCausalLM) (BitNetConfig 모델)
- **blenderbot** -- [BlenderbotForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/blenderbot#transformers.BlenderbotForCausalLM) (BlenderbotConfig 모델)
- **blenderbot-small** -- [BlenderbotSmallForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/blenderbot-small#transformers.BlenderbotSmallForCausalLM) (BlenderbotSmallConfig 모델)
- **bloom** -- [BloomForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomForCausalLM) (BloomConfig 모델)
- **blt** -- [BltForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/blt#transformers.BltForCausalLM) (BltConfig 모델)
- **camembert** -- [CamembertForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForCausalLM) (CamembertConfig 모델)
- **codegen** -- [CodeGenForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/codegen#transformers.CodeGenForCausalLM) (CodeGenConfig 모델)
- **cohere** -- [CohereForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cohere#transformers.CohereForCausalLM) (CohereConfig 모델)
- **cohere2** -- [Cohere2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cohere2#transformers.Cohere2ForCausalLM) (Cohere2Config 모델)
- **cohere2_moe** -- [Cohere2MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cohere2_moe#transformers.Cohere2MoeForCausalLM) (Cohere2MoeConfig 모델)
- **cohere_compass_text** -- [CohereCompassForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassForCausalLM) (CohereCompassTextConfig 모델)
- **cpmant** -- [CpmAntForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cpmant#transformers.CpmAntForCausalLM) (CpmAntConfig 모델)
- **ctrl** -- [CTRLLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLLMHeadModel) (CTRLConfig 모델)
- **cwm** -- [CwmForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/cwm#transformers.CwmForCausalLM) (CwmConfig 모델)
- **data2vec-text** -- [Data2VecTextForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForCausalLM) (Data2VecTextConfig 모델)
- **dbrx** -- [DbrxForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/dbrx#transformers.DbrxForCausalLM) (DbrxConfig 모델)
- **deepseek_v2** -- [DeepseekV2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v2#transformers.DeepseekV2ForCausalLM) (DeepseekV2Config 모델)
- **deepseek_v3** -- [DeepseekV3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3ForCausalLM) (DeepseekV3Config 모델)
- **deepseek_v32** -- [DeepseekV32ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v32#transformers.DeepseekV32ForCausalLM) (DeepseekV32Config 모델)
- **deepseek_v4** -- [DeepseekV4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v4#transformers.DeepseekV4ForCausalLM) (DeepseekV4Config 모델)
- **diffllama** -- [DiffLlamaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaForCausalLM) (DiffLlamaConfig 모델)
- **doge** -- [DogeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/doge#transformers.DogeForCausalLM) (DogeConfig 모델)
- **dots1** -- [Dots1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/dots1#transformers.Dots1ForCausalLM) (Dots1Config 모델)
- **dots3_note** -- [Dots3NoteForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteForCausalLM) (Dots3NoteConfig 모델)
- **dots3_note_text** -- [Dots3NoteForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteForCausalLM) (Dots3NoteTextConfig 모델)
- **electra** -- [ElectraForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForCausalLM) (ElectraConfig 모델)
- **emu3** -- [Emu3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3ForCausalLM) (Emu3Config 모델)
- **ernie** -- [ErnieForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForCausalLM) (ErnieConfig 모델)
- **ernie4_5** -- [Ernie4_5ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5#transformers.Ernie4_5ForCausalLM) (Ernie4_5Config 모델)
- **ernie4_5_moe** -- [Ernie4_5_MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_moe#transformers.Ernie4_5_MoeForCausalLM) (Ernie4_5_MoeConfig 모델)
- **exaone4** -- [Exaone4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4ForCausalLM) (Exaone4Config 모델)
- **exaone_moe** -- [ExaoneMoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/exaone_moe#transformers.ExaoneMoeForCausalLM) (ExaoneMoeConfig 모델)
- **falcon** -- [FalconForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconForCausalLM) (FalconConfig 모델)
- **falcon_h1** -- [FalconH1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/falcon_h1#transformers.FalconH1ForCausalLM) (FalconH1Config 모델)
- **falcon_mamba** -- [FalconMambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/falcon_mamba#transformers.FalconMambaForCausalLM) (FalconMambaConfig 모델)
- **flex_olmo** -- [FlexOlmoForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/flex_olmo#transformers.FlexOlmoForCausalLM) (FlexOlmoConfig 모델)
- **flint** -- [FlintForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/flint#transformers.FlintForCausalLM) (FlintConfig 모델)
- **fuyu** -- [FuyuForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuForCausalLM) (FuyuConfig 모델)
- **gemma** -- [GemmaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaForCausalLM) (GemmaConfig 모델)
- **gemma2** -- [Gemma2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2ForCausalLM) (Gemma2Config 모델)
- **gemma3** -- [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (Gemma3Config 모델)
- **gemma3_text** -- [Gemma3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForCausalLM) (Gemma3TextConfig 모델)
- **gemma3n** -- [Gemma3nForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nForConditionalGeneration) (Gemma3nConfig 모델)
- **gemma3n_text** -- [Gemma3nForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nForCausalLM) (Gemma3nTextConfig 모델)
- **gemma4** -- [Gemma4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4ForConditionalGeneration) (Gemma4Config 모델)
- **gemma4_assistant** -- [Gemma4AssistantForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma4_assistant#transformers.Gemma4AssistantForCausalLM) (Gemma4AssistantConfig 모델)
- **gemma4_text** -- [Gemma4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4ForCausalLM) (Gemma4TextConfig 모델)
- **gemma4_unified** -- [Gemma4UnifiedForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedForConditionalGeneration) (Gemma4UnifiedConfig 모델)
- **gemma4_unified_assistant** -- [Gemma4UnifiedAssistantForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified_assistant#transformers.Gemma4UnifiedAssistantForCausalLM) (Gemma4UnifiedAssistantConfig 모델)
- **gemma4_unified_text** -- [Gemma4UnifiedForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedForCausalLM) (Gemma4UnifiedTextConfig 모델)
- **git** -- [GitForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitForCausalLM) (GitConfig 모델)
- **glm** -- [GlmForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmForCausalLM) (GlmConfig 모델)
- **glm4** -- [Glm4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4ForCausalLM) (Glm4Config 모델)
- **glm4_moe** -- [Glm4MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe#transformers.Glm4MoeForCausalLM) (Glm4MoeConfig 모델)
- **glm4_moe_lite** -- [Glm4MoeLiteForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/glm4_moe_lite#transformers.Glm4MoeLiteForCausalLM) (Glm4MoeLiteConfig 모델)
- **glm_moe_dsa** -- [GlmMoeDsaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/glm_moe_dsa#transformers.GlmMoeDsaForCausalLM) (GlmMoeDsaConfig 모델)
- **got_ocr2** -- [GotOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2ForConditionalGeneration) (GotOcr2Config 모델)
- **gpt-sw3** -- [GPT2LMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2LMHeadModel) (GPT2Config 모델)
- **gpt2** -- [GPT2LMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2LMHeadModel) (GPT2Config 모델)
- **gpt_bigcode** -- [GPTBigCodeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeForCausalLM) (GPTBigCodeConfig 모델)
- **gpt_neo** -- [GPTNeoForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoForCausalLM) (GPTNeoConfig 모델)
- **gpt_neox** -- [GPTNeoXForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXForCausalLM) (GPTNeoXConfig 모델)
- **gpt_neox_japanese** -- [GPTNeoXJapaneseForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox_japanese#transformers.GPTNeoXJapaneseForCausalLM) (GPTNeoXJapaneseConfig 모델)
- **gpt_oss** -- [GptOssForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssForCausalLM) (GptOssConfig 모델)
- **gptj** -- [GPTJForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJForCausalLM) (GPTJConfig 모델)
- **granite** -- [GraniteForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granite#transformers.GraniteForCausalLM) (GraniteConfig 모델)
- **granite_swa** -- [GraniteSWAForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granite_swa#transformers.GraniteSWAForCausalLM) (GraniteSWAConfig 모델)
- **granitemoe** -- [GraniteMoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granitemoe#transformers.GraniteMoeForCausalLM) (GraniteMoeConfig 모델)
- **granitemoe_swa** -- [GraniteMoeSWAForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granitemoe_swa#transformers.GraniteMoeSWAForCausalLM) (GraniteMoeSWAConfig 모델)
- **granitemoehybrid** -- [GraniteMoeHybridForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granitemoehybrid#transformers.GraniteMoeHybridForCausalLM) (GraniteMoeHybridConfig 모델)
- **granitemoeshared** -- [GraniteMoeSharedForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/granitemoeshared#transformers.GraniteMoeSharedForCausalLM) (GraniteMoeSharedConfig 모델)
- **helium** -- [HeliumForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumForCausalLM) (HeliumConfig 모델)
- **hrm_text** -- [HrmTextForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hrm_text#transformers.HrmTextForCausalLM) (HrmTextConfig 모델)
- **hunyuan_v1_dense** -- [HunYuanDenseV1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_dense#transformers.HunYuanDenseV1ForCausalLM) (HunYuanDenseV1Config 모델)
- **hunyuan_v1_moe** -- [HunYuanMoEV1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_moe#transformers.HunYuanMoEV1ForCausalLM) (HunYuanMoEV1Config 모델)
- **hy_v3** -- [HYV3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hy_v3#transformers.HYV3ForCausalLM) (HYV3Config 모델)
- **hy_v4** -- [HYV4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hy_v4#transformers.HYV4ForCausalLM) (HYV4Config 모델)
- **hyperclovax** -- [HyperCLOVAXForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax#transformers.HyperCLOVAXForCausalLM) (HyperCLOVAXConfig 모델)
- **inkling2_text** -- [Inkling2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2ForCausalLM) (Inkling2TextConfig 모델)
- **inkling_text** -- [InklingForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingForCausalLM) (InklingTextConfig 모델)
- **internlm3** -- [LlamaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaForCausalLM) (InternLM3Config 모델)
- **iquestcoder** -- [LlamaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaForCausalLM) (IQuestCoderConfig 모델)
- **jais2** -- [Jais2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/jais2#transformers.Jais2ForCausalLM) (Jais2Config 모델)
- **jamba** -- [JambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/jamba#transformers.JambaForCausalLM) (JambaConfig 모델)
- **jetmoe** -- [JetMoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/jetmoe#transformers.JetMoeForCausalLM) (JetMoeConfig 모델)
- **kimi_linear** -- [KimiLinearForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/kimi_linear#transformers.KimiLinearForCausalLM) (KimiLinearConfig 모델)
- **kolibri1** -- [Kolibri1ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/kolibri1#transformers.Kolibri1ForCausalLM) (Kolibri1Config 모델)
- **laguna** -- [LagunaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/laguna#transformers.LagunaForCausalLM) (LagunaConfig 모델)
- **lfm2** -- [Lfm2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/lfm2#transformers.Lfm2ForCausalLM) (Lfm2Config 모델)
- **lfm2_moe** -- [Lfm2MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/lfm2_moe#transformers.Lfm2MoeForCausalLM) (Lfm2MoeConfig 모델)
- **lfm3** -- [Lfm3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/lfm3#transformers.Lfm3ForCausalLM) (Lfm3Config 모델)
- **llama** -- [LlamaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaForCausalLM) (LlamaConfig 모델)
- **llama4** -- [Llama4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4ForCausalLM) (Llama4Config 모델)
- **llama4_text** -- [Llama4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4ForCausalLM) (Llama4TextConfig 모델)
- **longcat_flash** -- [LongcatFlashForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/longcat_flash#transformers.LongcatFlashForCausalLM) (LongcatFlashConfig 모델)
- **mamba** -- [MambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mamba#transformers.MambaForCausalLM) (MambaConfig 모델)
- **mamba2** -- [Mamba2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mamba2#transformers.Mamba2ForCausalLM) (Mamba2Config 모델)
- **marian** -- [MarianForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/marian#transformers.MarianForCausalLM) (MarianConfig 모델)
- **mbart** -- [MBartForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartForCausalLM) (MBartConfig 모델)
- **megatron-bert** -- [MegatronBertForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForCausalLM) (MegatronBertConfig 모델)
- **mellum** -- [MellumForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mellum#transformers.MellumForCausalLM) (MellumConfig 모델)
- **mimo_v2_flash** -- [MiMoV2FlashForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mimo_v2_flash#transformers.MiMoV2FlashForCausalLM) (MiMoV2FlashConfig 모델)
- **minicpm3** -- [MiniCPM3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/minicpm3#transformers.MiniCPM3ForCausalLM) (MiniCPM3Config 모델)
- **minimax** -- [MiniMaxForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxForCausalLM) (MiniMaxConfig 모델)
- **minimax_m2** -- [MiniMaxM2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/minimax_m2#transformers.MiniMaxM2ForCausalLM) (MiniMaxM2Config 모델)
- **minimax_m3_vl_text** -- [MiniMaxM3VLForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLForCausalLM) (MiniMaxM3VLTextConfig 모델)
- **ministral** -- [MinistralForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralForCausalLM) (MinistralConfig 모델)
- **ministral3** -- [Ministral3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3ForCausalLM) (Ministral3Config 모델)
- **mistral** -- [MistralForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralForCausalLM) (MistralConfig 모델)
- **mixtral** -- [MixtralForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralForCausalLM) (MixtralConfig 모델)
- **mllama** -- [MllamaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaForCausalLM) (MllamaConfig 모델)
- **modernbert-decoder** -- [ModernBertDecoderForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/modernbert-decoder#transformers.ModernBertDecoderForCausalLM) (ModernBertDecoderConfig 모델)
- **moshi** -- [MoshiForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/moshi#transformers.MoshiForCausalLM) (MoshiConfig 모델)
- **moss_transcribe_diarize** -- [MossTranscribeDiarizeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeForConditionalGeneration) (MossTranscribeDiarizeConfig 모델)
- **mpt** -- [MptForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptForCausalLM) (MptConfig 모델)
- **musicgen** -- [MusicgenForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/musicgen#transformers.MusicgenForCausalLM) (MusicgenConfig 모델)
- **musicgen_melody** -- [MusicgenMelodyForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/musicgen_melody#transformers.MusicgenMelodyForCausalLM) (MusicgenMelodyConfig 모델)
- **mvp** -- [MvpForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForCausalLM) (MvpConfig 모델)
- **nanochat** -- [NanoChatForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/nanochat#transformers.NanoChatForCausalLM) (NanoChatConfig 모델)
- **nemotron** -- [NemotronForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronForCausalLM) (NemotronConfig 모델)
- **nemotron_h** -- [NemotronHForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h#transformers.NemotronHForCausalLM) (NemotronHConfig 모델)
- **olmo** -- [OlmoForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/olmo#transformers.OlmoForCausalLM) (OlmoConfig 모델)
- **olmo2** -- [Olmo2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/olmo2#transformers.Olmo2ForCausalLM) (Olmo2Config 모델)
- **olmo3** -- [Olmo3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/olmo3#transformers.Olmo3ForCausalLM) (Olmo3Config 모델)
- **olmo_hybrid** -- [OlmoHybridForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/olmo_hybrid#transformers.OlmoHybridForCausalLM) (OlmoHybridConfig 모델)
- **olmoe** -- [OlmoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/olmoe#transformers.OlmoeForCausalLM) (OlmoeConfig 모델)
- **openai-gpt** -- [OpenAIGPTLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTLMHeadModel) (OpenAIGPTConfig 모델)
- **opt** -- [OPTForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTForCausalLM) (OPTConfig 모델)
- **pegasus** -- [PegasusForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/pegasus#transformers.PegasusForCausalLM) (PegasusConfig 모델)
- **persimmon** -- [PersimmonForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonForCausalLM) (PersimmonConfig 모델)
- **phi** -- [PhiForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiForCausalLM) (PhiConfig 모델)
- **phi3** -- [Phi3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3ForCausalLM) (Phi3Config 모델)
- **phi4_multimodal** -- [Phi4MultimodalForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalForCausalLM) (Phi4MultimodalConfig 모델)
- **phimoe** -- [PhimoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/phimoe#transformers.PhimoeForCausalLM) (PhimoeConfig 모델)
- **plbart** -- [PLBartForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartForCausalLM) (PLBartConfig 모델)
- **prophetnet** -- [ProphetNetForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/prophetnet#transformers.ProphetNetForCausalLM) (ProphetNetConfig 모델)
- **qwen2** -- [Qwen2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2ForCausalLM) (Qwen2Config 모델)
- **qwen2_moe** -- [Qwen2MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeForCausalLM) (Qwen2MoeConfig 모델)
- **qwen3** -- [Qwen3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3ForCausalLM) (Qwen3Config 모델)
- **qwen3_5** -- [Qwen3_5ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForCausalLM) (Qwen3_5Config 모델)
- **qwen3_5_moe** -- [Qwen3_5MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeForCausalLM) (Qwen3_5MoeConfig 모델)
- **qwen3_5_moe_text** -- [Qwen3_5MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeForCausalLM) (Qwen3_5MoeTextConfig 모델)
- **qwen3_5_text** -- [Qwen3_5ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForCausalLM) (Qwen3_5TextConfig 모델)
- **qwen3_moe** -- [Qwen3MoeForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeForCausalLM) (Qwen3MoeConfig 모델)
- **qwen3_next** -- [Qwen3NextForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextForCausalLM) (Qwen3NextConfig 모델)
- **qwen4_exp** -- [Qwen4ExpForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpForCausalLM) (Qwen4ExpConfig 모델)
- **qwen4_exp_text** -- [Qwen4ExpForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpForCausalLM) (Qwen4ExpTextConfig 모델)
- **recurrent_gemma** -- [RecurrentGemmaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/recurrent_gemma#transformers.RecurrentGemmaForCausalLM) (RecurrentGemmaConfig 모델)
- **reformer** -- [ReformerModelWithLMHead](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerModelWithLMHead) (ReformerConfig 모델)
- **rembert** -- [RemBertForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForCausalLM) (RemBertConfig 모델)
- **roberta** -- [RobertaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForCausalLM) (RobertaConfig 모델)
- **roberta-prelayernorm** -- [RobertaPreLayerNormForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForCausalLM) (RobertaPreLayerNormConfig 모델)
- **roc_bert** -- [RoCBertForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForCausalLM) (RoCBertConfig 모델)
- **roformer** -- [RoFormerForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForCausalLM) (RoFormerConfig 모델)
- **rwkv** -- [RwkvForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/rwkv#transformers.RwkvForCausalLM) (RwkvConfig 모델)
- **seed_oss** -- [SeedOssForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssForCausalLM) (SeedOssConfig 모델)
- **smollm3** -- [SmolLM3ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3ForCausalLM) (SmolLM3Config 모델)
- **solar_open** -- [SolarOpenForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/solar_open#transformers.SolarOpenForCausalLM) (SolarOpenConfig 모델)
- **stablelm** -- [StableLmForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmForCausalLM) (StableLmConfig 모델)
- **starcoder2** -- [Starcoder2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2ForCausalLM) (Starcoder2Config 모델)
- **trocr** -- [TrOCRForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/trocr#transformers.TrOCRForCausalLM) (TrOCRConfig 모델)
- **vaultgemma** -- [VaultGemmaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/vaultgemma#transformers.VaultGemmaForCausalLM) (VaultGemmaConfig 모델)
- **whisper** -- [WhisperForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperForCausalLM) (WhisperConfig 모델)
- **xglm** -- [XGLMForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xglm#transformers.XGLMForCausalLM) (XGLMConfig 모델)
- **xlm** -- [XLMWithLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMWithLMHeadModel) (XLMConfig 모델)
- **xlm-roberta** -- [XLMRobertaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForCausalLM) (XLMRobertaConfig 모델)
- **xlm-roberta-xl** -- [XLMRobertaXLForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForCausalLM) (XLMRobertaXLConfig 모델)
- **xlnet** -- [XLNetLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetLMHeadModel) (XLNetConfig 모델)
- **xlstm** -- [xLSTMForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xlstm#transformers.xLSTMForCausalLM) (xLSTMConfig 모델)
- **xmod** -- [XmodForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForCausalLM) (XmodConfig 모델)
- **youtu** -- [YoutuForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/youtu#transformers.YoutuForCausalLM) (YoutuConfig 모델)
- **zamba** -- [ZambaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/zamba#transformers.ZambaForCausalLM) (ZambaConfig 모델)
- **zamba2** -- [Zamba2ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/zamba2#transformers.Zamba2ForCausalLM) (Zamba2Config 모델)
- **zaya** -- [ZayaForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/zaya#transformers.ZayaForCausalLM) (ZayaConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForCausalLM

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForCausalLM.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForCausalLM.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForMaskedLM[[transformers.AutoModelForMaskedLM]]

#### transformers.AutoModelForMaskedLM[[transformers.AutoModelForMaskedLM]]

```python
transformers.AutoModelForMaskedLM(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2364)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로(마스크 언어 모델링 헤드를 포함하여) 인스턴스화되는 범용 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForMaskedLM.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다: - [AlbertConfig](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertConfig) 구성 클래스: [AlbertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForMaskedLM) (AlbertConfig 모델) - [BartConfig](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartConfig) 구성 클래스: [BartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForConditionalGeneration) (BartConfig 모델) - [BertConfig](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertConfig) 구성 클래스: [BertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForMaskedLM) (BertConfig 모델) - [BigBirdConfig](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdConfig) 구성 클래스: [BigBirdForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForMaskedLM) (BigBirdConfig 모델) - [CamembertConfig](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertConfig) 구성 클래스: [CamembertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForMaskedLM) (CamembertConfig 모델) - [ConvBertConfig](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertConfig) 구성 클래스: [ConvBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertForMaskedLM) (ConvBertConfig 모델) - [Data2VecTextConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextConfig) 구성 클래스: [Data2VecTextForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForMaskedLM) (Data2VecTextConfig 모델) - [DebertaConfig](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaConfig) 구성 클래스: [DebertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaForMaskedLM) (DebertaConfig 모델) - [DebertaV2Config](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2Config) 구성 클래스: [DebertaV2ForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForMaskedLM) (DebertaV2Config 모델) - [DistilBertConfig](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertConfig) 구성 클래스: [DistilBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForMaskedLM) (DistilBertConfig 모델) - [ElectraConfig](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraConfig) 구성 클래스: [ElectraForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForMaskedLM) (ElectraConfig 모델) - [ErnieConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieConfig) 구성 클래스: [ErnieForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForMaskedLM) (ErnieConfig 모델) - [EsmConfig](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmConfig) 구성 클래스: [EsmForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmForMaskedLM) (EsmConfig 모델) - [EsmcConfig](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcConfig) 구성 클래스: [EsmcForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcForMaskedLM) (EsmcConfig 모델) - [EuroBertConfig](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertConfig) 구성 클래스: [EuroBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertForMaskedLM) (EuroBertConfig 모델) - [FNetConfig](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetConfig) 구성 클래스: [FNetForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForMaskedLM) (FNetConfig 모델) - [FlaubertConfig](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertConfig) 구성 클래스: [FlaubertWithLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertWithLMHeadModel) (FlaubertConfig 모델) - [FunnelConfig](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelConfig) 구성 클래스: [FunnelForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForMaskedLM) (FunnelConfig 모델) - [GteConfig](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteConfig) 구성 클래스: [GteForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteForMaskedLM) (GteConfig 모델) - [IBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertConfig) 구성 클래스: [IBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForMaskedLM) (IBertConfig 모델) - [JinaEmbeddingsV3Config](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3Config) 구성 클래스: [JinaEmbeddingsV3ForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3ForMaskedLM) (JinaEmbeddingsV3Config 모델) - [LayoutLMConfig](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMConfig) 구성 클래스: [LayoutLMForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMForMaskedLM) (LayoutLMConfig 모델) - [LongformerConfig](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerConfig) 구성 클래스: [LongformerForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForMaskedLM) (LongformerConfig 모델) - [LukeConfig](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeConfig) 구성 클래스: [LukeForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForMaskedLM) (LukeConfig 모델) - [MBartConfig](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartConfig) 구성 클래스: [MBartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartForConditionalGeneration) (MBartConfig 모델) - [MPNetConfig](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetConfig) 구성 클래스: [MPNetForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForMaskedLM) (MPNetConfig 모델) - [MegatronBertConfig](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertConfig) 구성 클래스: [MegatronBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForMaskedLM) (MegatronBertConfig 모델) - [MobileBertConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertConfig) 구성 클래스: [MobileBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForMaskedLM) (MobileBertConfig 모델) - [ModernBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertConfig) 구성 클래스: [ModernBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertForMaskedLM) (ModernBertConfig 모델) - [ModernVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertConfig) 구성 클래스: [ModernVBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertForMaskedLM) (ModernVBertConfig 모델) - [MraConfig](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraConfig) 구성 클래스: [MraForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForMaskedLM) (MraConfig 모델) - [MvpConfig](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpConfig) 구성 클래스: [MvpForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForConditionalGeneration) (MvpConfig 모델) - [NeoMMEConfig](/docs/transformers/v5.23.0/ko/model_doc/neomme#transformers.NeoMMEConfig) 구성 클래스: [NeoMMEForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/neomme#transformers.NeoMMEForMaskedLM) (NeoMMEConfig 모델) - [NomicBertConfig](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertConfig) 구성 클래스: [NomicBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertForMaskedLM) (NomicBertConfig 모델) - [NystromformerConfig](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerConfig) 구성 클래스: [NystromformerForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerForMaskedLM) (NystromformerConfig 모델) - [PerceiverConfig](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverConfig) 구성 클래스: [PerceiverForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverForMaskedLM) (PerceiverConfig 모델) - [ReformerConfig](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerConfig) 구성 클래스: [ReformerForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerForMaskedLM) (ReformerConfig 모델) - [RemBertConfig](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertConfig) 구성 클래스: [RemBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForMaskedLM) (RemBertConfig 모델) - [RoCBertConfig](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertConfig) 구성 클래스: [RoCBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForMaskedLM) (RoCBertConfig 모델) - [RoFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerConfig) 구성 클래스: [RoFormerForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForMaskedLM) (RoFormerConfig 모델) - [RobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaConfig) 구성 클래스: [RobertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForMaskedLM) (RobertaConfig 모델) - [RobertaPreLayerNormConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormConfig) 구성 클래스: [RobertaPreLayerNormForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForMaskedLM) (RobertaPreLayerNormConfig 모델) - [SqueezeBertConfig](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertConfig) 구성 클래스: [SqueezeBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForMaskedLM) (SqueezeBertConfig 모델) - [TapasConfig](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasConfig) 구성 클래스: [TapasForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasForMaskedLM) (TapasConfig 모델) - [UltraVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertConfig) 구성 클래스: [UltraVBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertForMaskedLM) (UltraVBertConfig 모델) - [UltrabertConfig](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertConfig) 구성 클래스: [UltrabertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertForMaskedLM) (UltrabertConfig 모델) - [XLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMConfig) 구성 클래스: [XLMWithLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMWithLMHeadModel) (XLMConfig 모델) - [XLMRobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaConfig) 구성 클래스: [XLMRobertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForMaskedLM) (XLMRobertaConfig 모델) - [XLMRobertaXLConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLConfig) 구성 클래스: [XLMRobertaXLForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForMaskedLM) (XLMRobertaXLConfig 모델) - [XmodConfig](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodConfig) 구성 클래스: [XmodForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForMaskedLM) (XmodConfig 모델) - [YosoConfig](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoConfig) 구성 클래스: [YosoForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoForMaskedLM) (YosoConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForMaskedLM

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForMaskedLM.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForMaskedLM.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **albert** -- [AlbertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForMaskedLM) (AlbertConfig 모델)
- **bart** -- [BartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForConditionalGeneration) (BartConfig 모델)
- **bert** -- [BertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForMaskedLM) (BertConfig 모델)
- **big_bird** -- [BigBirdForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForMaskedLM) (BigBirdConfig 모델)
- **camembert** -- [CamembertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForMaskedLM) (CamembertConfig 모델)
- **convbert** -- [ConvBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertForMaskedLM) (ConvBertConfig 모델)
- **data2vec-text** -- [Data2VecTextForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForMaskedLM) (Data2VecTextConfig 모델)
- **deberta** -- [DebertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaForMaskedLM) (DebertaConfig 모델)
- **deberta-v2** -- [DebertaV2ForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForMaskedLM) (DebertaV2Config 모델)
- **distilbert** -- [DistilBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForMaskedLM) (DistilBertConfig 모델)
- **electra** -- [ElectraForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForMaskedLM) (ElectraConfig 모델)
- **ernie** -- [ErnieForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForMaskedLM) (ErnieConfig 모델)
- **esm** -- [EsmForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmForMaskedLM) (EsmConfig 모델)
- **esmc** -- [EsmcForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcForMaskedLM) (EsmcConfig 모델)
- **eurobert** -- [EuroBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertForMaskedLM) (EuroBertConfig 모델)
- **flaubert** -- [FlaubertWithLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertWithLMHeadModel) (FlaubertConfig 모델)
- **fnet** -- [FNetForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForMaskedLM) (FNetConfig 모델)
- **funnel** -- [FunnelForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForMaskedLM) (FunnelConfig 모델)
- **gte** -- [GteForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteForMaskedLM) (GteConfig 모델)
- **ibert** -- [IBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForMaskedLM) (IBertConfig 모델)
- **jina_embeddings_v3** -- [JinaEmbeddingsV3ForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3ForMaskedLM) (JinaEmbeddingsV3Config 모델)
- **layoutlm** -- [LayoutLMForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMForMaskedLM) (LayoutLMConfig 모델)
- **longformer** -- [LongformerForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForMaskedLM) (LongformerConfig 모델)
- **luke** -- [LukeForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForMaskedLM) (LukeConfig 모델)
- **mbart** -- [MBartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartForConditionalGeneration) (MBartConfig 모델)
- **megatron-bert** -- [MegatronBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForMaskedLM) (MegatronBertConfig 모델)
- **mobilebert** -- [MobileBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForMaskedLM) (MobileBertConfig 모델)
- **modernbert** -- [ModernBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertForMaskedLM) (ModernBertConfig 모델)
- **modernvbert** -- [ModernVBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertForMaskedLM) (ModernVBertConfig 모델)
- **mpnet** -- [MPNetForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForMaskedLM) (MPNetConfig 모델)
- **mra** -- [MraForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForMaskedLM) (MraConfig 모델)
- **mvp** -- [MvpForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForConditionalGeneration) (MvpConfig 모델)
- **neomme** -- [NeoMMEForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/neomme#transformers.NeoMMEForMaskedLM) (NeoMMEConfig 모델)
- **nomic_bert** -- [NomicBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertForMaskedLM) (NomicBertConfig 모델)
- **nystromformer** -- [NystromformerForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerForMaskedLM) (NystromformerConfig 모델)
- **perceiver** -- [PerceiverForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverForMaskedLM) (PerceiverConfig 모델)
- **reformer** -- [ReformerForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerForMaskedLM) (ReformerConfig 모델)
- **rembert** -- [RemBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForMaskedLM) (RemBertConfig 모델)
- **roberta** -- [RobertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForMaskedLM) (RobertaConfig 모델)
- **roberta-prelayernorm** -- [RobertaPreLayerNormForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForMaskedLM) (RobertaPreLayerNormConfig 모델)
- **roc_bert** -- [RoCBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForMaskedLM) (RoCBertConfig 모델)
- **roformer** -- [RoFormerForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForMaskedLM) (RoFormerConfig 모델)
- **squeezebert** -- [SqueezeBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForMaskedLM) (SqueezeBertConfig 모델)
- **tapas** -- [TapasForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasForMaskedLM) (TapasConfig 모델)
- **ultrabert** -- [UltrabertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertForMaskedLM) (UltrabertConfig 모델)
- **ultravbert** -- [UltraVBertForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertForMaskedLM) (UltraVBertConfig 모델)
- **xlm** -- [XLMWithLMHeadModel](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMWithLMHeadModel) (XLMConfig 모델)
- **xlm-roberta** -- [XLMRobertaForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForMaskedLM) (XLMRobertaConfig 모델)
- **xlm-roberta-xl** -- [XLMRobertaXLForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForMaskedLM) (XLMRobertaXLConfig 모델)
- **xmod** -- [XmodForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForMaskedLM) (XmodConfig 모델)
- **yoso** -- [YosoForMaskedLM](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoForMaskedLM) (YosoConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForMaskedLM

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForMaskedLM.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForMaskedLM.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForMaskGeneration[[transformers.AutoModelForMaskGeneration]]

#### transformers.AutoModelForMaskGeneration[[transformers.AutoModelForMaskGeneration]]

```python
transformers.AutoModelForMaskGeneration(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2313)

### AutoModelForSeq2SeqLM[[transformers.AutoModelForSeq2SeqLM]]

#### transformers.AutoModelForSeq2SeqLM[[transformers.AutoModelForSeq2SeqLM]]

```python
transformers.AutoModelForSeq2SeqLM(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2371)

이것은 일반적인 모델 클래스로, [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때 라이브러리의 모델 클래스 중 하나로 인스턴스화됩니다(시퀀스-투-시퀀스 언어 모델링 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForSeq2SeqLM.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:  - [AudioFlamingo3Config](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3Config) 구성 클래스: [AudioFlamingo3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3ForConditionalGeneration) (AudioFlamingo3Config 모델) - [BartConfig](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartConfig) 구성 클래스: [BartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForConditionalGeneration) (BartConfig 모델) - [BigBirdPegasusConfig](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusConfig) 구성 클래스: [BigBirdPegasusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusForConditionalGeneration) (BigBirdPegasusConfig 모델) - [BlenderbotConfig](/docs/transformers/v5.23.0/ko/model_doc/blenderbot#transformers.BlenderbotConfig) 구성 클래스: [BlenderbotForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blenderbot#transformers.BlenderbotForConditionalGeneration) (BlenderbotConfig 모델) - [BlenderbotSmallConfig](/docs/transformers/v5.23.0/ko/model_doc/blenderbot-small#transformers.BlenderbotSmallConfig) 구성 클래스: [BlenderbotSmallForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blenderbot-small#transformers.BlenderbotSmallForConditionalGeneration) (BlenderbotSmallConfig 모델) - [EncoderDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/encoder-decoder#transformers.EncoderDecoderConfig) 구성 클래스: [EncoderDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/encoder-decoder#transformers.EncoderDecoderModel) (EncoderDecoderConfig 모델) - [FSMTConfig](/docs/transformers/v5.23.0/ko/model_doc/fsmt#transformers.FSMTConfig) 구성 클래스: [FSMTForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fsmt#transformers.FSMTForConditionalGeneration) (FSMTConfig 모델) - [GlmAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrConfig) 구성 클래스: [GlmAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrForConditionalGeneration) (GlmAsrConfig 모델) - [GraniteSpeechConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechConfig) 구성 클래스: [GraniteSpeechForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechForConditionalGeneration) (GraniteSpeechConfig 모델) - [GraniteSpeechPlusConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusConfig) 구성 클래스: [GraniteSpeechPlusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusForConditionalGeneration) (GraniteSpeechPlusConfig 모델) - [LEDConfig](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.LEDConfig) 구성 클래스: [LEDForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.LEDForConditionalGeneration) (LEDConfig 모델) - [LongT5Config](/docs/transformers/v5.23.0/ko/model_doc/longt5#transformers.LongT5Config) 구성 클래스: [LongT5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/longt5#transformers.LongT5ForConditionalGeneration) (LongT5Config 모델) - [M2M100Config](/docs/transformers/v5.23.0/ko/model_doc/m2m_100#transformers.M2M100Config) 구성 클래스: [M2M100ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/m2m_100#transformers.M2M100ForConditionalGeneration) (M2M100Config 모델) - [MBartConfig](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartConfig) 구성 클래스: [MBartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartForConditionalGeneration) (MBartConfig 모델) - [MT5Config](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5Config) 구성 클래스: [MT5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5ForConditionalGeneration) (MT5Config 모델) - [MarianConfig](/docs/transformers/v5.23.0/ko/model_doc/marian#transformers.MarianConfig) 구성 클래스: [MarianMTModel](/docs/transformers/v5.23.0/ko/model_doc/marian#transformers.MarianMTModel) (MarianConfig 모델) - [MossTranscribeDiarizeConfig](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeConfig) 구성 클래스: [MossTranscribeDiarizeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeForConditionalGeneration) (MossTranscribeDiarizeConfig 모델) - [MusicFlamingoConfig](/docs/transformers/v5.23.0/ko/model_doc/musicflamingo#transformers.MusicFlamingoConfig) 구성 클래스: [MusicFlamingoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/musicflamingo#transformers.MusicFlamingoForConditionalGeneration) (MusicFlamingoConfig 모델) - [MvpConfig](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpConfig) 구성 클래스: [MvpForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForConditionalGeneration) (MvpConfig 모델) - [Nemotron3_5TranscribeConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeConfig) 구성 클래스: [Nemotron3_5TranscribeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeForConditionalGeneration) (Nemotron3_5TranscribeConfig 모델) - [NllbMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/nllb-moe#transformers.NllbMoeConfig) 구성 클래스: [NllbMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/nllb-moe#transformers.NllbMoeForConditionalGeneration) (NllbMoeConfig 모델) - [PLBartConfig](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartConfig) 구성 클래스: [PLBartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartForConditionalGeneration) (PLBartConfig 모델) - [PegasusConfig](/docs/transformers/v5.23.0/ko/model_doc/pegasus#transformers.PegasusConfig) 구성 클래스: [PegasusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pegasus#transformers.PegasusForConditionalGeneration) (PegasusConfig 모델) - [PegasusXConfig](/docs/transformers/v5.23.0/ko/model_doc/pegasus_x#transformers.PegasusXConfig) 구성 클래스: [PegasusXForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pegasus_x#transformers.PegasusXForConditionalGeneration) (PegasusXConfig 모델) - [ProphetNetConfig](/docs/transformers/v5.23.0/ko/model_doc/prophetnet#transformers.ProphetNetConfig) 구성 클래스: [ProphetNetForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/prophetnet#transformers.ProphetNetForConditionalGeneration) (ProphetNetConfig 모델) - [Qwen2AudioConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioConfig) 구성 클래스: [Qwen2AudioForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioForConditionalGeneration) (Qwen2AudioConfig 모델) - [Qwen3ASRConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRConfig) 구성 클래스: [Qwen3ASRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRForConditionalGeneration) (Qwen3ASRConfig 모델) - [SeamlessM4TConfig](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TConfig) 구성 클래스: [SeamlessM4TForTextToText](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TForTextToText) (SeamlessM4TConfig 모델) - [SeamlessM4Tv2Config](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t_v2#transformers.SeamlessM4Tv2Config) 구성 클래스: [SeamlessM4Tv2ForTextToText](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t_v2#transformers.SeamlessM4Tv2ForTextToText) (SeamlessM4Tv2Config 모델) - [SwitchTransformersConfig](/docs/transformers/v5.23.0/ko/model_doc/switch_transformers#transformers.SwitchTransformersConfig) 구성 클래스: [SwitchTransformersForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/switch_transformers#transformers.SwitchTransformersForConditionalGeneration) (SwitchTransformersConfig 모델) - [T5Config](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Config) 구성 클래스: [T5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5ForConditionalGeneration) (T5Config 모델) - [T5Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2Config) 구성 클래스: [T5Gemma2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForConditionalGeneration) (T5Gemma2Config 모델) - [T5GemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaConfig) 구성 클래스: [T5GemmaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaForConditionalGeneration) (T5GemmaConfig 모델) - [UMT5Config](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5Config) 구성 클래스: [UMT5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5ForConditionalGeneration) (UMT5Config 모델) - [VibeVoiceAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrConfig) 구성 클래스: [VibeVoiceAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrForConditionalGeneration) (VibeVoiceAsrConfig 모델) - [VoxtralConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralConfig) 구성 클래스: [VoxtralForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralForConditionalGeneration) (VoxtralConfig 모델) - [VoxtralRealtimeConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeConfig) 구성 클래스: [VoxtralRealtimeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeForConditionalGeneration) (VoxtralRealtimeConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForSeq2SeqLM

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-t5/t5-base")
>>> model = AutoModelForSeq2SeqLM.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForSeq2SeqLM.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **audioflamingo3** -- [AudioFlamingo3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/audioflamingo3#transformers.AudioFlamingo3ForConditionalGeneration) (AudioFlamingo3Config 모델)
- **bart** -- [BartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForConditionalGeneration) (BartConfig 모델)
- **bigbird_pegasus** -- [BigBirdPegasusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusForConditionalGeneration) (BigBirdPegasusConfig 모델)
- **blenderbot** -- [BlenderbotForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blenderbot#transformers.BlenderbotForConditionalGeneration) (BlenderbotConfig 모델)
- **blenderbot-small** -- [BlenderbotSmallForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blenderbot-small#transformers.BlenderbotSmallForConditionalGeneration) (BlenderbotSmallConfig 모델)
- **인코더-디코더** -- [EncoderDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/encoder-decoder#transformers.EncoderDecoderModel) (EncoderDecoderConfig 모델)
- **fsmt** -- [FSMTForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fsmt#transformers.FSMTForConditionalGeneration) (FSMTConfig 모델)
- **glmasr** -- [GlmAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrForConditionalGeneration) (GlmAsrConfig 모델)
- **granite_speech** -- [GraniteSpeechForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechForConditionalGeneration) (GraniteSpeechConfig 모델)
- **granite_speech_plus** -- [GraniteSpeechPlusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusForConditionalGeneration) (GraniteSpeechPlusConfig 모델)
- **led** -- [LEDForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.LEDForConditionalGeneration) (LEDConfig 모델)
- **longt5** -- [LongT5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/longt5#transformers.LongT5ForConditionalGeneration) (LongT5Config 모델)
- **m2m_100** -- [M2M100ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/m2m_100#transformers.M2M100ForConditionalGeneration) (M2M100Config 모델)
- **marian** -- [MarianMTModel](/docs/transformers/v5.23.0/ko/model_doc/marian#transformers.MarianMTModel) (MarianConfig 모델)
- **mbart** -- [MBartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartForConditionalGeneration) (MBartConfig 모델)
- **moss_transcribe_diarize** -- [MossTranscribeDiarizeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeForConditionalGeneration) (MossTranscribeDiarizeConfig 모델)
- **mt5** -- [MT5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5ForConditionalGeneration) (MT5Config 모델)
- **musicflamingo** -- [MusicFlamingoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/musicflamingo#transformers.MusicFlamingoForConditionalGeneration) (MusicFlamingoConfig 모델)
- **mvp** -- [MvpForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForConditionalGeneration) (MvpConfig 모델)
- **nemo_speechlm** -- [Nemotron3_5TranscribeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeForConditionalGeneration) (Nemotron3_5TranscribeConfig 모델)
- **nllb-moe** -- [NllbMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/nllb-moe#transformers.NllbMoeForConditionalGeneration) (NllbMoeConfig 모델)
- **pegasus** -- [PegasusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pegasus#transformers.PegasusForConditionalGeneration) (PegasusConfig 모델)
- **pegasus_x** -- [PegasusXForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pegasus_x#transformers.PegasusXForConditionalGeneration) (PegasusXConfig 모델)
- **plbart** -- [PLBartForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartForConditionalGeneration) (PLBartConfig 모델)
- **prophetnet** -- [ProphetNetForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/prophetnet#transformers.ProphetNetForConditionalGeneration) (ProphetNetConfig 모델)
- **qwen2_audio** -- [Qwen2AudioForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioForConditionalGeneration) (Qwen2AudioConfig 모델)
- **qwen3_asr** -- [Qwen3ASRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRForConditionalGeneration) (Qwen3ASRConfig 모델)
- **seamless_m4t** -- [SeamlessM4TForTextToText](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TForTextToText) (SeamlessM4TConfig 모델)
- **seamless_m4t_v2** -- [SeamlessM4Tv2ForTextToText](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t_v2#transformers.SeamlessM4Tv2ForTextToText) (SeamlessM4Tv2Config 모델)
- **switch_transformers** -- [SwitchTransformersForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/switch_transformers#transformers.SwitchTransformersForConditionalGeneration) (SwitchTransformersConfig 모델)
- **t5** -- [T5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5ForConditionalGeneration) (T5Config 모델)
- **t5gemma** -- [T5GemmaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaForConditionalGeneration) (T5GemmaConfig 모델)
- **t5gemma2** -- [T5Gemma2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForConditionalGeneration) (T5Gemma2Config 모델)
- **umt5** -- [UMT5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5ForConditionalGeneration) (UMT5Config 모델)
- **vibevoice_asr** -- [VibeVoiceAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrForConditionalGeneration) (VibeVoiceAsrConfig 모델)
- **voxtral** -- [VoxtralForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralForConditionalGeneration) (VoxtralConfig 모델)
- **voxtral_realtime** -- [VoxtralRealtimeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeForConditionalGeneration) (VoxtralRealtimeConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForSeq2SeqLM

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForSeq2SeqLM.from_pretrained("google-t5/t5-base")

>>> # Update configuration during loading
>>> model = AutoModelForSeq2SeqLM.from_pretrained("google-t5/t5-base", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForSequenceClassification[[transformers.AutoModelForSequenceClassification]]

#### transformers.AutoModelForSequenceClassification[[transformers.AutoModelForSequenceClassification]]

```python
transformers.AutoModelForSequenceClassification(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2382)

이것은 범용 모델 클래스로, [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때 라이브러리의 모델 클래스 중 하나(시퀀스 분류 헤드를 포함)로 인스턴스화됩니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForSequenceClassification.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : The model class to instantiate is selected based on the configuration class:  - [AXK1Config](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1Config) configuration class: [AXK1ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1ForSequenceClassification) (AXK1Config model) - [AXK2Config](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2Config) configuration class: [AXK2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2ForSequenceClassification) (AXK2Config model) - [AlbertConfig](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertConfig) configuration class: [AlbertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForSequenceClassification) (AlbertConfig model) - [ArceeConfig](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeConfig) configuration class: [ArceeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeForSequenceClassification) (ArceeConfig model) - [BartConfig](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartConfig) configuration class: [BartForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForSequenceClassification) (BartConfig model) - [BertConfig](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertConfig) configuration class: [BertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForSequenceClassification) (BertConfig model) - [BigBirdConfig](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdConfig) configuration class: [BigBirdForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForSequenceClassification) (BigBirdConfig model) - [BigBirdPegasusConfig](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusConfig) configuration class: [BigBirdPegasusForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusForSequenceClassification) (BigBirdPegasusConfig model) - [BioGptConfig](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptConfig) configuration class: [BioGptForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptForSequenceClassification) (BioGptConfig model) - [BloomConfig](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomConfig) configuration class: [BloomForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomForSequenceClassification) (BloomConfig model) - [CTRLConfig](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLConfig) configuration class: [CTRLForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLForSequenceClassification) (CTRLConfig model) - [CamembertConfig](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertConfig) configuration class: [CamembertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForSequenceClassification) (CamembertConfig model) - [CanineConfig](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineConfig) configuration class: [CanineForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineForSequenceClassification) (CanineConfig model) - [CohereCompassTextConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassTextConfig) configuration class: [CohereCompassTextForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassTextForSequenceClassification) (CohereCompassTextConfig model) - [ConvBertConfig](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertConfig) configuration class: [ConvBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertForSequenceClassification) (ConvBertConfig model) - [Data2VecTextConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextConfig) configuration class: [Data2VecTextForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForSequenceClassification) (Data2VecTextConfig model) - [DebertaConfig](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaConfig) configuration class: [DebertaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaForSequenceClassification) (DebertaConfig model) - [DebertaV2Config](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2Config) configuration class: [DebertaV2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForSequenceClassification) (DebertaV2Config model) - [DeepseekV2Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v2#transformers.DeepseekV2Config) configuration class: [DeepseekV2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v2#transformers.DeepseekV2ForSequenceClassification) (DeepseekV2Config model) - [DeepseekV3Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3Config) configuration class: [DeepseekV3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3ForSequenceClassification) (DeepseekV3Config model) - [DiffLlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaConfig) configuration class: [DiffLlamaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaForSequenceClassification) (DiffLlamaConfig model) - [DistilBertConfig](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertConfig) configuration class: [DistilBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForSequenceClassification) (DistilBertConfig model) - [DogeConfig](/docs/transformers/v5.23.0/ko/model_doc/doge#transformers.DogeConfig) configuration class: [DogeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/doge#transformers.DogeForSequenceClassification) (DogeConfig model) - [ElectraConfig](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraConfig) configuration class: [ElectraForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForSequenceClassification) (ElectraConfig model) - [ErnieConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieConfig) configuration class: [ErnieForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForSequenceClassification) (ErnieConfig model) - [EsmConfig](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmConfig) configuration class: [EsmForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmForSequenceClassification) (EsmConfig model) - [EsmcConfig](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcConfig) configuration class: [EsmcForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcForSequenceClassification) (EsmcConfig model) - [EuroBertConfig](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertConfig) configuration class: [EuroBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertForSequenceClassification) (EuroBertConfig model) - [Exaone4Config](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4Config) configuration class: [Exaone4ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4ForSequenceClassification) (Exaone4Config model) - [FNetConfig](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetConfig) configuration class: [FNetForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForSequenceClassification) (FNetConfig model) - [FalconConfig](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconConfig) configuration class: [FalconForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconForSequenceClassification) (FalconConfig model) - [FlaubertConfig](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertConfig) configuration class: [FlaubertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertForSequenceClassification) (FlaubertConfig model) - [FunnelConfig](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelConfig) configuration class: [FunnelForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForSequenceClassification) (FunnelConfig model) - [GPT2Config](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Config) configuration class: [GPT2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2ForSequenceClassification) (GPT2Config model) - [GPTBigCodeConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeConfig) configuration class: [GPTBigCodeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeForSequenceClassification) (GPTBigCodeConfig model) - [GPTJConfig](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJConfig) configuration class: [GPTJForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJForSequenceClassification) (GPTJConfig model) - [GPTNeoConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoConfig) configuration class: [GPTNeoForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoForSequenceClassification) (GPTNeoConfig model) - [GPTNeoXConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXConfig) configuration class: [GPTNeoXForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXForSequenceClassification) (GPTNeoXConfig model) - [Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2Config) configuration class: [Gemma2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2ForSequenceClassification) (Gemma2Config model) - [Gemma3Config](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Config) configuration class: [Gemma3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForSequenceClassification) (Gemma3Config model) - [Gemma3TextConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3TextConfig) configuration class: [Gemma3TextForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3TextForSequenceClassification) (Gemma3TextConfig model) - [GemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaConfig) configuration class: [GemmaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaForSequenceClassification) (GemmaConfig model) - [Glm4Config](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4Config) configuration class: [Glm4ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4ForSequenceClassification) (Glm4Config model) - [GlmConfig](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmConfig) configuration class: [GlmForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmForSequenceClassification) (GlmConfig model) - [GptOssConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssConfig) configuration class: [GptOssForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssForSequenceClassification) (GptOssConfig model) - [GteConfig](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteConfig) configuration class: [GteForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteForSequenceClassification) (GteConfig model) - [HeliumConfig](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumConfig) configuration class: [HeliumForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumForSequenceClassification) (HeliumConfig model) - [HunYuanDenseV1Config](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_dense#transformers.HunYuanDenseV1Config) configuration class: [HunYuanDenseV1ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_dense#transformers.HunYuanDenseV1ForSequenceClassification) (HunYuanDenseV1Config model) - [HunYuanMoEV1Config](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_moe#transformers.HunYuanMoEV1Config) configuration class: [HunYuanMoEV1ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_moe#transformers.HunYuanMoEV1ForSequenceClassification) (HunYuanMoEV1Config model) - [IBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertConfig) configuration class: [IBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForSequenceClassification) (IBertConfig model) - [JambaConfig](/docs/transformers/v5.23.0/ko/model_doc/jamba#transformers.JambaConfig) configuration class: [JambaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/jamba#transformers.JambaForSequenceClassification) (JambaConfig model) - [JetMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/jetmoe#transformers.JetMoeConfig) configuration class: [JetMoeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/jetmoe#transformers.JetMoeForSequenceClassification) (JetMoeConfig model) - [JinaEmbeddingsV3Config](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3Config) configuration class: [JinaEmbeddingsV3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3ForSequenceClassification) (JinaEmbeddingsV3Config model) - [LayoutLMConfig](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMConfig) configuration class: [LayoutLMForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMForSequenceClassification) (LayoutLMConfig model) - [LayoutLMv2Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2Config) configuration class: [LayoutLMv2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2ForSequenceClassification) (LayoutLMv2Config model) - [LayoutLMv3Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3Config) configuration class: [LayoutLMv3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3ForSequenceClassification) (LayoutLMv3Config model) - [LiltConfig](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltConfig) configuration class: [LiltForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltForSequenceClassification) (LiltConfig model) - [LlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaConfig) configuration class: [LlamaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaForSequenceClassification) (LlamaConfig model) - [LongformerConfig](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerConfig) configuration class: [LongformerForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForSequenceClassification) (LongformerConfig model) - [LukeConfig](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeConfig) configuration class: [LukeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForSequenceClassification) (LukeConfig model) - [MBartConfig](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartConfig) configuration class: [MBartForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartForSequenceClassification) (MBartConfig model) - [MPNetConfig](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetConfig) configuration class: [MPNetForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForSequenceClassification) (MPNetConfig model) - [MT5Config](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5Config) configuration class: [MT5ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5ForSequenceClassification) (MT5Config model) - [MarkupLMConfig](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMConfig) configuration class: [MarkupLMForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMForSequenceClassification) (MarkupLMConfig model) - [MegatronBertConfig](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertConfig) configuration class: [MegatronBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForSequenceClassification) (MegatronBertConfig model) - [MiniCPM3Config](/docs/transformers/v5.23.0/ko/model_doc/minicpm3#transformers.MiniCPM3Config) configuration class: [MiniCPM3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/minicpm3#transformers.MiniCPM3ForSequenceClassification) (MiniCPM3Config model) - [MiniMaxConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxConfig) configuration class: [MiniMaxForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxForSequenceClassification) (MiniMaxConfig model) - [Ministral3Config](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3Config) configuration class: [Ministral3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3ForSequenceClassification) (Ministral3Config model) - [MinistralConfig](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralConfig) configuration class: [MinistralForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralForSequenceClassification) (MinistralConfig model) - [Mistral4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4Config) configuration class: [Mistral4ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4ForSequenceClassification) (Mistral4Config model) - [MistralConfig](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralConfig) configuration class: [MistralForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralForSequenceClassification) (MistralConfig model) - [MixtralConfig](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralConfig) configuration class: [MixtralForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralForSequenceClassification) (MixtralConfig model) - [MobileBertConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertConfig) configuration class: [MobileBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForSequenceClassification) (MobileBertConfig model) - [ModernBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertConfig) configuration class: [ModernBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertForSequenceClassification) (ModernBertConfig model) - [ModernBertDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/modernbert-decoder#transformers.ModernBertDecoderConfig) configuration class: [ModernBertDecoderForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/modernbert-decoder#transformers.ModernBertDecoderForSequenceClassification) (ModernBertDecoderConfig model) - [ModernVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertConfig) configuration class: [ModernVBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertForSequenceClassification) (ModernVBertConfig model) - [MptConfig](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptConfig) configuration class: [MptForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptForSequenceClassification) (MptConfig model) - [MraConfig](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraConfig) configuration class: [MraForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForSequenceClassification) (MraConfig model) - [MvpConfig](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpConfig) configuration class: [MvpForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForSequenceClassification) (MvpConfig model) - [NemotronConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronConfig) configuration class: [NemotronForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronForSequenceClassification) (NemotronConfig model) - [NomicBertConfig](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertConfig) configuration class: [NomicBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertForSequenceClassification) (NomicBertConfig model) - [NystromformerConfig](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerConfig) configuration class: [NystromformerForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerForSequenceClassification) (NystromformerConfig model) - [OPTConfig](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTConfig) configuration class: [OPTForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTForSequenceClassification) (OPTConfig model) - [Olmo2Config](/docs/transformers/v5.23.0/ko/model_doc/olmo2#transformers.Olmo2Config) configuration class: [Olmo2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/olmo2#transformers.Olmo2ForSequenceClassification) (Olmo2Config model) - [Olmo3Config](/docs/transformers/v5.23.0/ko/model_doc/olmo3#transformers.Olmo3Config) configuration class: [Olmo3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/olmo3#transformers.Olmo3ForSequenceClassification) (Olmo3Config model) - [OlmoConfig](/docs/transformers/v5.23.0/ko/model_doc/olmo#transformers.OlmoConfig) configuration class: [OlmoForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/olmo#transformers.OlmoForSequenceClassification) (OlmoConfig model) - [OpenAIGPTConfig](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTConfig) configuration class: [OpenAIGPTForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTForSequenceClassification) (OpenAIGPTConfig model) - [PLBartConfig](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartConfig) configuration class: [PLBartForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartForSequenceClassification) (PLBartConfig model) - [PerceiverConfig](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverConfig) configuration class: [PerceiverForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverForSequenceClassification) (PerceiverConfig model) - [PersimmonConfig](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonConfig) configuration class: [PersimmonForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonForSequenceClassification) (PersimmonConfig model) - [Phi3Config](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3Config) configuration class: [Phi3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3ForSequenceClassification) (Phi3Config model) - [PhiConfig](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiConfig) configuration class: [PhiForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiForSequenceClassification) (PhiConfig model) - [PhimoeConfig](/docs/transformers/v5.23.0/ko/model_doc/phimoe#transformers.PhimoeConfig) configuration class: [PhimoeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/phimoe#transformers.PhimoeForSequenceClassification) (PhimoeConfig model) - [Qwen2Config](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Config) configuration class: [Qwen2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2ForSequenceClassification) (Qwen2Config model) - [Qwen2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeConfig) configuration class: [Qwen2MoeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeForSequenceClassification) (Qwen2MoeConfig model) - [Qwen3Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3Config) configuration class: [Qwen3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3ForSequenceClassification) (Qwen3Config model) - [Qwen3MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeConfig) configuration class: [Qwen3MoeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeForSequenceClassification) (Qwen3MoeConfig model) - [Qwen3NextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextConfig) configuration class: [Qwen3NextForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextForSequenceClassification) (Qwen3NextConfig model) - [Qwen3_5Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Config) configuration class: [Qwen3_5ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForSequenceClassification) (Qwen3_5Config model) - [Qwen3_5TextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5TextConfig) configuration class: [Qwen3_5TextForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5TextForSequenceClassification) (Qwen3_5TextConfig model) - [ReformerConfig](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerConfig) configuration class: [ReformerForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerForSequenceClassification) (ReformerConfig model) - [RemBertConfig](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertConfig) configuration class: [RemBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForSequenceClassification) (RemBertConfig model) - [RoCBertConfig](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertConfig) configuration class: [RoCBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForSequenceClassification) (RoCBertConfig model) - [RoFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerConfig) configuration class: [RoFormerForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForSequenceClassification) (RoFormerConfig model) - [RobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaConfig) configuration class: [RobertaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForSequenceClassification) (RobertaConfig model) - [RobertaPreLayerNormConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormConfig) configuration class: [RobertaPreLayerNormForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForSequenceClassification) (RobertaPreLayerNormConfig model) - [SeedOssConfig](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssConfig) configuration class: [SeedOssForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssForSequenceClassification) (SeedOssConfig model) - [SmolLM3Config](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3Config) configuration class: [SmolLM3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3ForSequenceClassification) (SmolLM3Config model) - [SqueezeBertConfig](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertConfig) configuration class: [SqueezeBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForSequenceClassification) (SqueezeBertConfig model) - [StableLmConfig](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmConfig) configuration class: [StableLmForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmForSequenceClassification) (StableLmConfig model) - [Starcoder2Config](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2Config) configuration class: [Starcoder2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2ForSequenceClassification) (Starcoder2Config model) - [T5Config](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Config) configuration class: [T5ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5ForSequenceClassification) (T5Config model) - [T5Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2Config) configuration class: [T5Gemma2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForSequenceClassification) (T5Gemma2Config model) - [T5GemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaConfig) configuration class: [T5GemmaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaForSequenceClassification) (T5GemmaConfig model) - [TapasConfig](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasConfig) configuration class: [TapasForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasForSequenceClassification) (TapasConfig model) - [UMT5Config](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5Config) configuration class: [UMT5ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5ForSequenceClassification) (UMT5Config model) - [UltraVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertConfig) configuration class: [UltraVBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertForSequenceClassification) (UltraVBertConfig model) - [UltrabertConfig](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertConfig) configuration class: [UltrabertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertForSequenceClassification) (UltrabertConfig model) - [XLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMConfig) configuration class: [XLMForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMForSequenceClassification) (XLMConfig model) - [XLMRobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaConfig) configuration class: [XLMRobertaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForSequenceClassification) (XLMRobertaConfig model) - [XLMRobertaXLConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLConfig) configuration class: [XLMRobertaXLForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForSequenceClassification) (XLMRobertaXLConfig model) - [XLNetConfig](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetConfig) configuration class: [XLNetForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetForSequenceClassification) (XLNetConfig model) - [XmodConfig](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodConfig) configuration class: [XmodForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForSequenceClassification) (XmodConfig model) - [YosoConfig](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoConfig) configuration class: [YosoForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoForSequenceClassification) (YosoConfig model) - [Zamba2Config](/docs/transformers/v5.23.0/ko/model_doc/zamba2#transformers.Zamba2Config) configuration class: [Zamba2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/zamba2#transformers.Zamba2ForSequenceClassification) (Zamba2Config model) - [ZambaConfig](/docs/transformers/v5.23.0/ko/model_doc/zamba#transformers.ZambaConfig) configuration class: [ZambaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/zamba#transformers.ZambaForSequenceClassification) (ZambaConfig model)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

configuration에서 라이브러리의 모델 클래스 중 하나(시퀀스 분류 헤드를 포함)를 초기화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForSequenceClassification

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForSequenceClassification.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForSequenceClassification.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **albert** -- [AlbertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForSequenceClassification) (AlbertConfig 모델)
- **arcee** -- [ArceeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeForSequenceClassification) (ArceeConfig 모델)
- **axk1** -- [AXK1ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1ForSequenceClassification) (AXK1Config 모델)
- **axk2** -- [AXK2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2ForSequenceClassification) (AXK2Config 모델)
- **bart** -- [BartForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForSequenceClassification) (BartConfig 모델)
- **bert** -- [BertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForSequenceClassification) (BertConfig 모델)
- **big_bird** -- [BigBirdForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForSequenceClassification) (BigBirdConfig 모델)
- **bigbird_pegasus** -- [BigBirdPegasusForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusForSequenceClassification) (BigBirdPegasusConfig 모델)
- **biogpt** -- [BioGptForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptForSequenceClassification) (BioGptConfig 모델)
- **bloom** -- [BloomForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomForSequenceClassification) (BloomConfig 모델)
- **camembert** -- [CamembertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForSequenceClassification) (CamembertConfig 모델)
- **canine** -- [CanineForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineForSequenceClassification) (CanineConfig 모델)
- **cohere_compass_text** -- [CohereCompassTextForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassTextForSequenceClassification) (CohereCompassTextConfig 모델)
- **convbert** -- [ConvBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertForSequenceClassification) (ConvBertConfig 모델)
- **ctrl** -- [CTRLForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ctrl#transformers.CTRLForSequenceClassification) (CTRLConfig 모델)
- **data2vec-text** -- [Data2VecTextForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForSequenceClassification) (Data2VecTextConfig 모델)
- **deberta** -- [DebertaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaForSequenceClassification) (DebertaConfig 모델)
- **deberta-v2** -- [DebertaV2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForSequenceClassification) (DebertaV2Config 모델)
- **deepseek_v2** -- [DeepseekV2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v2#transformers.DeepseekV2ForSequenceClassification) (DeepseekV2Config 모델)
- **deepseek_v3** -- [DeepseekV3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3ForSequenceClassification) (DeepseekV3Config 모델)
- **diffllama** -- [DiffLlamaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaForSequenceClassification) (DiffLlamaConfig 모델)
- **distilbert** -- [DistilBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForSequenceClassification) (DistilBertConfig 모델)
- **doge** -- [DogeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/doge#transformers.DogeForSequenceClassification) (DogeConfig 모델)
- **electra** -- [ElectraForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForSequenceClassification) (ElectraConfig 모델)
- **ernie** -- [ErnieForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForSequenceClassification) (ErnieConfig 모델)
- **esm** -- [EsmForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmForSequenceClassification) (EsmConfig 모델)
- **esmc** -- [EsmcForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcForSequenceClassification) (EsmcConfig 모델)
- **eurobert** -- [EuroBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertForSequenceClassification) (EuroBertConfig 모델)
- **exaone4** -- [Exaone4ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4ForSequenceClassification) (Exaone4Config 모델)
- **falcon** -- [FalconForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconForSequenceClassification) (FalconConfig 모델)
- **flaubert** -- [FlaubertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertForSequenceClassification) (FlaubertConfig 모델)
- **fnet** -- [FNetForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForSequenceClassification) (FNetConfig 모델)
- **funnel** -- [FunnelForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForSequenceClassification) (FunnelConfig 모델)
- **gemma** -- [GemmaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaForSequenceClassification) (GemmaConfig 모델)
- **gemma2** -- [Gemma2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2ForSequenceClassification) (Gemma2Config 모델)
- **gemma3** -- [Gemma3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForSequenceClassification) (Gemma3Config 모델)
- **gemma3_text** -- [Gemma3TextForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3TextForSequenceClassification) (Gemma3TextConfig 모델)
- **glm** -- [GlmForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmForSequenceClassification) (GlmConfig 모델)
- **glm4** -- [Glm4ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4ForSequenceClassification) (Glm4Config 모델)
- **gpt-sw3** -- [GPT2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2ForSequenceClassification) (GPT2Config 모델)
- **gpt2** -- [GPT2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2ForSequenceClassification) (GPT2Config 모델)
- **gpt_bigcode** -- [GPTBigCodeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeForSequenceClassification) (GPTBigCodeConfig 모델)
- **gpt_neo** -- [GPTNeoForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoForSequenceClassification) (GPTNeoConfig 모델)
- **gpt_neox** -- [GPTNeoXForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXForSequenceClassification) (GPTNeoXConfig 모델)
- **gpt_oss** -- [GptOssForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssForSequenceClassification) (GptOssConfig 모델)
- **gptj** -- [GPTJForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJForSequenceClassification) (GPTJConfig 모델)
- **gte** -- [GteForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteForSequenceClassification) (GteConfig 모델)
- **helium** -- [HeliumForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumForSequenceClassification) (HeliumConfig 모델)
- **hunyuan_v1_dense** -- [HunYuanDenseV1ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_dense#transformers.HunYuanDenseV1ForSequenceClassification) (HunYuanDenseV1Config 모델)
- **hunyuan_v1_moe** -- [HunYuanMoEV1ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_v1_moe#transformers.HunYuanMoEV1ForSequenceClassification) (HunYuanMoEV1Config 모델)
- **ibert** -- [IBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForSequenceClassification) (IBertConfig 모델)
- **jamba** -- [JambaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/jamba#transformers.JambaForSequenceClassification) (JambaConfig 모델)
- **jetmoe** -- [JetMoeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/jetmoe#transformers.JetMoeForSequenceClassification) (JetMoeConfig 모델)
- **jina_embeddings_v3** -- [JinaEmbeddingsV3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3ForSequenceClassification) (JinaEmbeddingsV3Config 모델)
- **layoutlm** -- [LayoutLMForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMForSequenceClassification) (LayoutLMConfig 모델)
- **layoutlmv2** -- [LayoutLMv2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2ForSequenceClassification) (LayoutLMv2Config 모델)
- **layoutlmv3** -- [LayoutLMv3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3ForSequenceClassification) (LayoutLMv3Config 모델)
- **lilt** -- [LiltForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltForSequenceClassification) (LiltConfig 모델)
- **llama** -- [LlamaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaForSequenceClassification) (LlamaConfig 모델)
- **longformer** -- [LongformerForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForSequenceClassification) (LongformerConfig 모델)
- **luke** -- [LukeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForSequenceClassification) (LukeConfig 모델)
- **markuplm** -- [MarkupLMForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMForSequenceClassification) (MarkupLMConfig 모델)
- **mbart** -- [MBartForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartForSequenceClassification) (MBartConfig 모델)
- **megatron-bert** -- [MegatronBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForSequenceClassification) (MegatronBertConfig 모델)
- **minicpm3** -- [MiniCPM3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/minicpm3#transformers.MiniCPM3ForSequenceClassification) (MiniCPM3Config 모델)
- **minimax** -- [MiniMaxForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxForSequenceClassification) (MiniMaxConfig 모델)
- **ministral** -- [MinistralForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralForSequenceClassification) (MinistralConfig 모델)
- **ministral3** -- [Ministral3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3ForSequenceClassification) (Ministral3Config 모델)
- **mistral** -- [MistralForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralForSequenceClassification) (MistralConfig 모델)
- **mistral4** -- [Mistral4ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4ForSequenceClassification) (Mistral4Config 모델)
- **mixtral** -- [MixtralForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralForSequenceClassification) (MixtralConfig 모델)
- **mobilebert** -- [MobileBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForSequenceClassification) (MobileBertConfig 모델)
- **modernbert** -- [ModernBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertForSequenceClassification) (ModernBertConfig 모델)
- **modernbert-decoder** -- [ModernBertDecoderForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/modernbert-decoder#transformers.ModernBertDecoderForSequenceClassification) (ModernBertDecoderConfig 모델)
- **modernvbert** -- [ModernVBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertForSequenceClassification) (ModernVBertConfig 모델)
- **mpnet** -- [MPNetForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForSequenceClassification) (MPNetConfig 모델)
- **mpt** -- [MptForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptForSequenceClassification) (MptConfig 모델)
- **mra** -- [MraForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForSequenceClassification) (MraConfig 모델)
- **mt5** -- [MT5ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5ForSequenceClassification) (MT5Config 모델)
- **mvp** -- [MvpForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForSequenceClassification) (MvpConfig 모델)
- **nemotron** -- [NemotronForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronForSequenceClassification) (NemotronConfig 모델)
- **nomic_bert** -- [NomicBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertForSequenceClassification) (NomicBertConfig 모델)
- **nystromformer** -- [NystromformerForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerForSequenceClassification) (NystromformerConfig 모델)
- **olmo** -- [OlmoForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/olmo#transformers.OlmoForSequenceClassification) (OlmoConfig 모델)
- **olmo2** -- [Olmo2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/olmo2#transformers.Olmo2ForSequenceClassification) (Olmo2Config 모델)
- **olmo3** -- [Olmo3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/olmo3#transformers.Olmo3ForSequenceClassification) (Olmo3Config 모델)
- **openai-gpt** -- [OpenAIGPTForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/openai-gpt#transformers.OpenAIGPTForSequenceClassification) (OpenAIGPTConfig 모델)
- **opt** -- [OPTForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTForSequenceClassification) (OPTConfig 모델)
- **perceiver** -- [PerceiverForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverForSequenceClassification) (PerceiverConfig 모델)
- **persimmon** -- [PersimmonForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonForSequenceClassification) (PersimmonConfig 모델)
- **phi** -- [PhiForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiForSequenceClassification) (PhiConfig 모델)
- **phi3** -- [Phi3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3ForSequenceClassification) (Phi3Config 모델)
- **phimoe** -- [PhimoeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/phimoe#transformers.PhimoeForSequenceClassification) (PhimoeConfig 모델)
- **plbart** -- [PLBartForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/plbart#transformers.PLBartForSequenceClassification) (PLBartConfig 모델)
- **qwen2** -- [Qwen2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2ForSequenceClassification) (Qwen2Config 모델)
- **qwen2_moe** -- [Qwen2MoeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeForSequenceClassification) (Qwen2MoeConfig 모델)
- **qwen3** -- [Qwen3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3ForSequenceClassification) (Qwen3Config 모델)
- **qwen3_5** -- [Qwen3_5ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForSequenceClassification) (Qwen3_5Config 모델)
- **qwen3_5_text** -- [Qwen3_5TextForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5TextForSequenceClassification) (Qwen3_5TextConfig 모델)
- **qwen3_moe** -- [Qwen3MoeForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeForSequenceClassification) (Qwen3MoeConfig 모델)
- **qwen3_next** -- [Qwen3NextForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextForSequenceClassification) (Qwen3NextConfig 모델)
- **reformer** -- [ReformerForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerForSequenceClassification) (ReformerConfig 모델)
- **rembert** -- [RemBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForSequenceClassification) (RemBertConfig 모델)
- **roberta** -- [RobertaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForSequenceClassification) (RobertaConfig 모델)
- **roberta-prelayernorm** -- [RobertaPreLayerNormForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForSequenceClassification) (RobertaPreLayerNormConfig 모델)
- **roc_bert** -- [RoCBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForSequenceClassification) (RoCBertConfig 모델)
- **roformer** -- [RoFormerForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForSequenceClassification) (RoFormerConfig 모델)
- **seed_oss** -- [SeedOssForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssForSequenceClassification) (SeedOssConfig 모델)
- **smollm3** -- [SmolLM3ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3ForSequenceClassification) (SmolLM3Config 모델)
- **squeezebert** -- [SqueezeBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForSequenceClassification) (SqueezeBertConfig 모델)
- **stablelm** -- [StableLmForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmForSequenceClassification) (StableLmConfig 모델)
- **starcoder2** -- [Starcoder2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2ForSequenceClassification) (Starcoder2Config 모델)
- **t5** -- [T5ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5ForSequenceClassification) (T5Config 모델)
- **t5gemma** -- [T5GemmaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaForSequenceClassification) (T5GemmaConfig 모델)
- **t5gemma2** -- [T5Gemma2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForSequenceClassification) (T5Gemma2Config 모델)
- **tapas** -- [TapasForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasForSequenceClassification) (TapasConfig 모델)
- **ultrabert** -- [UltrabertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertForSequenceClassification) (UltrabertConfig 모델)
- **ultravbert** -- [UltraVBertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertForSequenceClassification) (UltraVBertConfig 모델)
- **umt5** -- [UMT5ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5ForSequenceClassification) (UMT5Config 모델)
- **xlm** -- [XLMForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMForSequenceClassification) (XLMConfig 모델)
- **xlm-roberta** -- [XLMRobertaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForSequenceClassification) (XLMRobertaConfig 모델)
- **xlm-roberta-xl** -- [XLMRobertaXLForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForSequenceClassification) (XLMRobertaXLConfig 모델)
- **xlnet** -- [XLNetForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetForSequenceClassification) (XLNetConfig 모델)
- **xmod** -- [XmodForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForSequenceClassification) (XmodConfig 모델)
- **yoso** -- [YosoForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoForSequenceClassification) (YosoConfig 모델)
- **zamba** -- [ZambaForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/zamba#transformers.ZambaForSequenceClassification) (ZambaConfig 모델)
- **zamba2** -- [Zamba2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/zamba2#transformers.Zamba2ForSequenceClassification) (Zamba2Config 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForSequenceClassification

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForSequenceClassification.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForSequenceClassification.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForMultipleChoice[[transformers.AutoModelForMultipleChoice]]

#### transformers.AutoModelForMultipleChoice[[transformers.AutoModelForMultipleChoice]]

```python
transformers.AutoModelForMultipleChoice(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2438)

이것은 범용 모델 클래스로, [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때 라이브러리의 모델 클래스 중 하나(다중 선택 헤드를 포함)로 인스턴스화됩니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForMultipleChoice.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다: - [AlbertConfig](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertConfig) 구성 클래스: [AlbertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForMultipleChoice) (AlbertConfig 모델) - [BertConfig](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertConfig) 구성 클래스: [BertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForMultipleChoice) (BertConfig 모델) - [BigBirdConfig](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdConfig) 구성 클래스: [BigBirdForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForMultipleChoice) (BigBirdConfig 모델) - [CamembertConfig](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertConfig) 구성 클래스: [CamembertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForMultipleChoice) (CamembertConfig 모델) - [CanineConfig](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineConfig) 구성 클래스: [CanineForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineForMultipleChoice) (CanineConfig 모델) - [ConvBertConfig](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertConfig) 구성 클래스: [ConvBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertForMultipleChoice) (ConvBertConfig 모델) - [Data2VecTextConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextConfig) 구성 클래스: [Data2VecTextForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForMultipleChoice) (Data2VecTextConfig 모델) - [DebertaV2Config](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2Config) 구성 클래스: [DebertaV2ForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForMultipleChoice) (DebertaV2Config 모델) - [DistilBertConfig](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertConfig) 구성 클래스: [DistilBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForMultipleChoice) (DistilBertConfig 모델) - [ElectraConfig](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraConfig) 구성 클래스: [ElectraForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForMultipleChoice) (ElectraConfig 모델) - [ErnieConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieConfig) 구성 클래스: [ErnieForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForMultipleChoice) (ErnieConfig 모델) - [FNetConfig](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetConfig) 구성 클래스: [FNetForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForMultipleChoice) (FNetConfig 모델) - [FlaubertConfig](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertConfig) 구성 클래스: [FlaubertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertForMultipleChoice) (FlaubertConfig 모델) - [FunnelConfig](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelConfig) 구성 클래스: [FunnelForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForMultipleChoice) (FunnelConfig 모델) - [IBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertConfig) 구성 클래스: [IBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForMultipleChoice) (IBertConfig 모델) - [LongformerConfig](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerConfig) 구성 클래스: [LongformerForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForMultipleChoice) (LongformerConfig 모델) - [LukeConfig](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeConfig) 구성 클래스: [LukeForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForMultipleChoice) (LukeConfig 모델) - [MPNetConfig](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetConfig) 구성 클래스: [MPNetForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForMultipleChoice) (MPNetConfig 모델) - [MegatronBertConfig](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertConfig) 구성 클래스: [MegatronBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForMultipleChoice) (MegatronBertConfig 모델) - [MobileBertConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertConfig) 구성 클래스: [MobileBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForMultipleChoice) (MobileBertConfig 모델) - [ModernBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertConfig) 구성 클래스: [ModernBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertForMultipleChoice) (ModernBertConfig 모델) - [MraConfig](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraConfig) 구성 클래스: [MraForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForMultipleChoice) (MraConfig 모델) - [NystromformerConfig](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerConfig) 구성 클래스: [NystromformerForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerForMultipleChoice) (NystromformerConfig 모델) - [RemBertConfig](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertConfig) 구성 클래스: [RemBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForMultipleChoice) (RemBertConfig 모델) - [RoCBertConfig](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertConfig) 구성 클래스: [RoCBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForMultipleChoice) (RoCBertConfig 모델) - [RoFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerConfig) 구성 클래스: [RoFormerForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForMultipleChoice) (RoFormerConfig 모델) - [RobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaConfig) 구성 클래스: [RobertaForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForMultipleChoice) (RobertaConfig 모델) - [RobertaPreLayerNormConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormConfig) 구성 클래스: [RobertaPreLayerNormForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForMultipleChoice) (RobertaPreLayerNormConfig 모델) - [SqueezeBertConfig](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertConfig) 구성 클래스: [SqueezeBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForMultipleChoice) (SqueezeBertConfig 모델) - [XLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMConfig) 구성 클래스: [XLMForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMForMultipleChoice) (XLMConfig 모델) - [XLMRobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaConfig) 구성 클래스: [XLMRobertaForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForMultipleChoice) (XLMRobertaConfig 모델) - [XLMRobertaXLConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLConfig) 구성 클래스: [XLMRobertaXLForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForMultipleChoice) (XLMRobertaXLConfig 모델) - [XLNetConfig](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetConfig) 구성 클래스: [XLNetForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetForMultipleChoice) (XLNetConfig 모델) - [XmodConfig](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodConfig) 구성 클래스: [XmodForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForMultipleChoice) (XmodConfig 모델) - [YosoConfig](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoConfig) 구성 클래스: [YosoForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoForMultipleChoice) (YosoConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

configuration에서 라이브러리의 모델 클래스 중 하나(다중 선택 헤드를 포함)를 초기화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForMultipleChoice

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForMultipleChoice.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForMultipleChoice.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **albert** -- [AlbertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForMultipleChoice) (AlbertConfig 모델)
- **bert** -- [BertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForMultipleChoice) (BertConfig 모델)
- **big_bird** -- [BigBirdForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForMultipleChoice) (BigBirdConfig 모델)
- **camembert** -- [CamembertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForMultipleChoice) (CamembertConfig 모델)
- **canine** -- [CanineForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineForMultipleChoice) (CanineConfig 모델)
- **convbert** -- [ConvBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertForMultipleChoice) (ConvBertConfig 모델)
- **data2vec-text** -- [Data2VecTextForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForMultipleChoice) (Data2VecTextConfig 모델)
- **deberta-v2** -- [DebertaV2ForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForMultipleChoice) (DebertaV2Config 모델)
- **distilbert** -- [DistilBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForMultipleChoice) (DistilBertConfig 모델)
- **electra** -- [ElectraForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForMultipleChoice) (ElectraConfig 모델)
- **ernie** -- [ErnieForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForMultipleChoice) (ErnieConfig 모델)
- **flaubert** -- [FlaubertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertForMultipleChoice) (FlaubertConfig 모델)
- **fnet** -- [FNetForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForMultipleChoice) (FNetConfig 모델)
- **funnel** -- [FunnelForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForMultipleChoice) (FunnelConfig 모델)
- **ibert** -- [IBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForMultipleChoice) (IBertConfig 모델)
- **longformer** -- [LongformerForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForMultipleChoice) (LongformerConfig 모델)
- **luke** -- [LukeForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForMultipleChoice) (LukeConfig 모델)
- **megatron-bert** -- [MegatronBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForMultipleChoice) (MegatronBertConfig 모델)
- **mobilebert** -- [MobileBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForMultipleChoice) (MobileBertConfig 모델)
- **modernbert** -- [ModernBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertForMultipleChoice) (ModernBertConfig 모델)
- **mpnet** -- [MPNetForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForMultipleChoice) (MPNetConfig 모델)
- **mra** -- [MraForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForMultipleChoice) (MraConfig 모델)
- **nystromformer** -- [NystromformerForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerForMultipleChoice) (NystromformerConfig 모델)
- **rembert** -- [RemBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForMultipleChoice) (RemBertConfig 모델)
- **roberta** -- [RobertaForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForMultipleChoice) (RobertaConfig 모델)
- **roberta-prelayernorm** -- [RobertaPreLayerNormForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForMultipleChoice) (RobertaPreLayerNormConfig 모델)
- **roc_bert** -- [RoCBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForMultipleChoice) (RoCBertConfig 모델)
- **roformer** -- [RoFormerForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForMultipleChoice) (RoFormerConfig 모델)
- **squeezebert** -- [SqueezeBertForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForMultipleChoice) (SqueezeBertConfig 모델)
- **xlm** -- [XLMForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMForMultipleChoice) (XLMConfig 모델)
- **xlm-roberta** -- [XLMRobertaForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForMultipleChoice) (XLMRobertaConfig 모델)
- **xlm-roberta-xl** -- [XLMRobertaXLForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForMultipleChoice) (XLMRobertaXLConfig 모델)
- **xlnet** -- [XLNetForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetForMultipleChoice) (XLNetConfig 모델)
- **xmod** -- [XmodForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForMultipleChoice) (XmodConfig 모델)
- **yoso** -- [YosoForMultipleChoice](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoForMultipleChoice) (YosoConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForMultipleChoice

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForMultipleChoice.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForMultipleChoice.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForNextSentencePrediction[[transformers.AutoModelForNextSentencePrediction]]

#### transformers.AutoModelForNextSentencePrediction[[transformers.AutoModelForNextSentencePrediction]]

```python
transformers.AutoModelForNextSentencePrediction(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2445)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로(다음 문장 예측 헤드를 포함하여) 인스턴스화되는 범용 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForNextSentencePrediction.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다: - [BertConfig](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertConfig) 구성 클래스: [BertForNextSentencePrediction](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForNextSentencePrediction) (BertConfig 모델) - [ErnieConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieConfig) 구성 클래스: [ErnieForNextSentencePrediction](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForNextSentencePrediction) (ErnieConfig 모델) - [FNetConfig](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetConfig) 구성 클래스: [FNetForNextSentencePrediction](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForNextSentencePrediction) (FNetConfig 모델) - [MegatronBertConfig](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertConfig) 구성 클래스: [MegatronBertForNextSentencePrediction](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForNextSentencePrediction) (MegatronBertConfig 모델) - [MobileBertConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertConfig) 구성 클래스: [MobileBertForNextSentencePrediction](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForNextSentencePrediction) (MobileBertConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForNextSentencePrediction

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForNextSentencePrediction.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForNextSentencePrediction.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **bert** -- [BertForNextSentencePrediction](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForNextSentencePrediction) (BertConfig 모델)
- **ernie** -- [ErnieForNextSentencePrediction](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForNextSentencePrediction) (ErnieConfig 모델)
- **fnet** -- [FNetForNextSentencePrediction](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForNextSentencePrediction) (FNetConfig 모델)
- **megatron-bert** -- [MegatronBertForNextSentencePrediction](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForNextSentencePrediction) (MegatronBertConfig 모델)
- **mobilebert** -- [MobileBertForNextSentencePrediction](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForNextSentencePrediction) (MobileBertConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForNextSentencePrediction

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForNextSentencePrediction.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForNextSentencePrediction.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForTokenClassification[[transformers.AutoModelForTokenClassification]]

#### transformers.AutoModelForTokenClassification[[transformers.AutoModelForTokenClassification]]

```python
transformers.AutoModelForTokenClassification(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2431)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 범용 모델 클래스입니다(토큰 분류 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForTokenClassification.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : The model class to instantiate is selected based on the configuration class:  - [AXK1Config](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1Config) configuration class: [AXK1ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1ForTokenClassification) (AXK1Config model) - [AXK2Config](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2Config) configuration class: [AXK2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2ForTokenClassification) (AXK2Config model) - [AlbertConfig](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertConfig) configuration class: [AlbertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForTokenClassification) (AlbertConfig model) - [ApertusConfig](/docs/transformers/v5.23.0/ko/model_doc/apertus#transformers.ApertusConfig) configuration class: [ApertusForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/apertus#transformers.ApertusForTokenClassification) (ApertusConfig model) - [ArceeConfig](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeConfig) configuration class: [ArceeForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeForTokenClassification) (ArceeConfig model) - [BertConfig](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertConfig) configuration class: [BertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForTokenClassification) (BertConfig model) - [BigBirdConfig](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdConfig) configuration class: [BigBirdForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForTokenClassification) (BigBirdConfig model) - [BioGptConfig](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptConfig) configuration class: [BioGptForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptForTokenClassification) (BioGptConfig model) - [BloomConfig](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomConfig) configuration class: [BloomForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomForTokenClassification) (BloomConfig model) - [BrosConfig](/docs/transformers/v5.23.0/ko/model_doc/bros#transformers.BrosConfig) configuration class: [BrosForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/bros#transformers.BrosForTokenClassification) (BrosConfig model) - [CamembertConfig](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertConfig) configuration class: [CamembertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForTokenClassification) (CamembertConfig model) - [CanineConfig](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineConfig) configuration class: [CanineForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineForTokenClassification) (CanineConfig model) - [ConvBertConfig](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertConfig) configuration class: [ConvBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertForTokenClassification) (ConvBertConfig model) - [Data2VecTextConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextConfig) configuration class: [Data2VecTextForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForTokenClassification) (Data2VecTextConfig model) - [DebertaConfig](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaConfig) configuration class: [DebertaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaForTokenClassification) (DebertaConfig model) - [DebertaV2Config](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2Config) configuration class: [DebertaV2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForTokenClassification) (DebertaV2Config model) - [DeepseekV3Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3Config) configuration class: [DeepseekV3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3ForTokenClassification) (DeepseekV3Config model) - [DiffLlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaConfig) configuration class: [DiffLlamaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaForTokenClassification) (DiffLlamaConfig model) - [DistilBertConfig](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertConfig) configuration class: [DistilBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForTokenClassification) (DistilBertConfig model) - [ElectraConfig](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraConfig) configuration class: [ElectraForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForTokenClassification) (ElectraConfig model) - [ErnieConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieConfig) configuration class: [ErnieForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForTokenClassification) (ErnieConfig model) - [EsmConfig](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmConfig) configuration class: [EsmForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmForTokenClassification) (EsmConfig model) - [EsmcConfig](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcConfig) configuration class: [EsmcForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcForTokenClassification) (EsmcConfig model) - [EuroBertConfig](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertConfig) configuration class: [EuroBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertForTokenClassification) (EuroBertConfig model) - [Exaone4Config](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4Config) configuration class: [Exaone4ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4ForTokenClassification) (Exaone4Config model) - [FNetConfig](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetConfig) configuration class: [FNetForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForTokenClassification) (FNetConfig model) - [FalconConfig](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconConfig) configuration class: [FalconForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconForTokenClassification) (FalconConfig model) - [FlaubertConfig](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertConfig) configuration class: [FlaubertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertForTokenClassification) (FlaubertConfig model) - [FunnelConfig](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelConfig) configuration class: [FunnelForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForTokenClassification) (FunnelConfig model) - [GPT2Config](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Config) configuration class: [GPT2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2ForTokenClassification) (GPT2Config model) - [GPTBigCodeConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeConfig) configuration class: [GPTBigCodeForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeForTokenClassification) (GPTBigCodeConfig model) - [GPTNeoConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoConfig) configuration class: [GPTNeoForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoForTokenClassification) (GPTNeoConfig model) - [GPTNeoXConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXConfig) configuration class: [GPTNeoXForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXForTokenClassification) (GPTNeoXConfig model) - [Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2Config) configuration class: [Gemma2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2ForTokenClassification) (Gemma2Config model) - [GemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaConfig) configuration class: [GemmaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaForTokenClassification) (GemmaConfig model) - [Glm4Config](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4Config) configuration class: [Glm4ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4ForTokenClassification) (Glm4Config model) - [GlmConfig](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmConfig) configuration class: [GlmForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmForTokenClassification) (GlmConfig model) - [GptOssConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssConfig) configuration class: [GptOssForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssForTokenClassification) (GptOssConfig model) - [GteConfig](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteConfig) configuration class: [GteForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteForTokenClassification) (GteConfig model) - [HeliumConfig](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumConfig) configuration class: [HeliumForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumForTokenClassification) (HeliumConfig model) - [IBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertConfig) configuration class: [IBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForTokenClassification) (IBertConfig model) - [JinaEmbeddingsV3Config](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3Config) configuration class: [JinaEmbeddingsV3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3ForTokenClassification) (JinaEmbeddingsV3Config model) - [LayoutLMConfig](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMConfig) configuration class: [LayoutLMForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMForTokenClassification) (LayoutLMConfig model) - [LayoutLMv2Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2Config) configuration class: [LayoutLMv2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2ForTokenClassification) (LayoutLMv2Config model) - [LayoutLMv3Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3Config) configuration class: [LayoutLMv3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3ForTokenClassification) (LayoutLMv3Config model) - [LiltConfig](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltConfig) configuration class: [LiltForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltForTokenClassification) (LiltConfig model) - [LlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaConfig) configuration class: [LlamaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/llama#transformers.LlamaForTokenClassification) (LlamaConfig model) - [LongformerConfig](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerConfig) configuration class: [LongformerForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForTokenClassification) (LongformerConfig model) - [LukeConfig](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeConfig) configuration class: [LukeForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForTokenClassification) (LukeConfig model) - [MPNetConfig](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetConfig) configuration class: [MPNetForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForTokenClassification) (MPNetConfig model) - [MT5Config](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5Config) configuration class: [MT5ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5ForTokenClassification) (MT5Config model) - [MarkupLMConfig](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMConfig) configuration class: [MarkupLMForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMForTokenClassification) (MarkupLMConfig model) - [MegatronBertConfig](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertConfig) configuration class: [MegatronBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForTokenClassification) (MegatronBertConfig model) - [MiniMaxConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxConfig) configuration class: [MiniMaxForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxForTokenClassification) (MiniMaxConfig model) - [Ministral3Config](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3Config) configuration class: [Ministral3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3ForTokenClassification) (Ministral3Config model) - [MinistralConfig](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralConfig) configuration class: [MinistralForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralForTokenClassification) (MinistralConfig model) - [Mistral4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4Config) configuration class: [Mistral4ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4ForTokenClassification) (Mistral4Config model) - [MistralConfig](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralConfig) configuration class: [MistralForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralForTokenClassification) (MistralConfig model) - [MixtralConfig](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralConfig) configuration class: [MixtralForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralForTokenClassification) (MixtralConfig model) - [MobileBertConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertConfig) configuration class: [MobileBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForTokenClassification) (MobileBertConfig model) - [ModernBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertConfig) configuration class: [ModernBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertForTokenClassification) (ModernBertConfig model) - [ModernVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertConfig) configuration class: [ModernVBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertForTokenClassification) (ModernVBertConfig model) - [MptConfig](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptConfig) configuration class: [MptForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptForTokenClassification) (MptConfig model) - [MraConfig](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraConfig) configuration class: [MraForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForTokenClassification) (MraConfig model) - [NemotronConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronConfig) configuration class: [NemotronForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronForTokenClassification) (NemotronConfig model) - [NomicBertConfig](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertConfig) configuration class: [NomicBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertForTokenClassification) (NomicBertConfig model) - [NystromformerConfig](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerConfig) configuration class: [NystromformerForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerForTokenClassification) (NystromformerConfig model) - [OpenAIPrivacyFilterConfig](/docs/transformers/v5.23.0/ko/model_doc/openai_privacy_filter#transformers.OpenAIPrivacyFilterConfig) configuration class: [OpenAIPrivacyFilterForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/openai_privacy_filter#transformers.OpenAIPrivacyFilterForTokenClassification) (OpenAIPrivacyFilterConfig model) - [PersimmonConfig](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonConfig) configuration class: [PersimmonForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonForTokenClassification) (PersimmonConfig model) - [Phi3Config](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3Config) configuration class: [Phi3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3ForTokenClassification) (Phi3Config model) - [PhiConfig](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiConfig) configuration class: [PhiForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiForTokenClassification) (PhiConfig model) - [Qwen2Config](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Config) configuration class: [Qwen2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2ForTokenClassification) (Qwen2Config model) - [Qwen2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeConfig) configuration class: [Qwen2MoeForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeForTokenClassification) (Qwen2MoeConfig model) - [Qwen3ASRConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRConfig) configuration class: [Qwen3ASRForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRForTokenClassification) (Qwen3ASRConfig model) - [Qwen3Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3Config) configuration class: [Qwen3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3ForTokenClassification) (Qwen3Config model) - [Qwen3MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeConfig) configuration class: [Qwen3MoeForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeForTokenClassification) (Qwen3MoeConfig model) - [Qwen3NextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextConfig) configuration class: [Qwen3NextForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextForTokenClassification) (Qwen3NextConfig model) - [Qwen3_5Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Config) configuration class: [Qwen3_5ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForTokenClassification) (Qwen3_5Config model) - [RemBertConfig](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertConfig) configuration class: [RemBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForTokenClassification) (RemBertConfig model) - [RoCBertConfig](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertConfig) configuration class: [RoCBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForTokenClassification) (RoCBertConfig model) - [RoFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerConfig) configuration class: [RoFormerForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForTokenClassification) (RoFormerConfig model) - [RobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaConfig) configuration class: [RobertaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForTokenClassification) (RobertaConfig model) - [RobertaPreLayerNormConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormConfig) configuration class: [RobertaPreLayerNormForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForTokenClassification) (RobertaPreLayerNormConfig model) - [SeedOssConfig](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssConfig) configuration class: [SeedOssForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssForTokenClassification) (SeedOssConfig model) - [SmolLM3Config](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3Config) configuration class: [SmolLM3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3ForTokenClassification) (SmolLM3Config model) - [SqueezeBertConfig](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertConfig) configuration class: [SqueezeBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForTokenClassification) (SqueezeBertConfig model) - [StableLmConfig](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmConfig) configuration class: [StableLmForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmForTokenClassification) (StableLmConfig model) - [Starcoder2Config](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2Config) configuration class: [Starcoder2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2ForTokenClassification) (Starcoder2Config model) - [T5Config](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Config) configuration class: [T5ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5ForTokenClassification) (T5Config model) - [T5Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2Config) configuration class: [T5Gemma2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForTokenClassification) (T5Gemma2Config model) - [T5GemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaConfig) configuration class: [T5GemmaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaForTokenClassification) (T5GemmaConfig model) - [UMT5Config](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5Config) configuration class: [UMT5ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5ForTokenClassification) (UMT5Config model) - [UltraVBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertConfig) configuration class: [UltraVBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertForTokenClassification) (UltraVBertConfig model) - [UltrabertConfig](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertConfig) configuration class: [UltrabertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertForTokenClassification) (UltrabertConfig model) - [XLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMConfig) configuration class: [XLMForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMForTokenClassification) (XLMConfig model) - [XLMRobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaConfig) configuration class: [XLMRobertaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForTokenClassification) (XLMRobertaConfig model) - [XLMRobertaXLConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLConfig) configuration class: [XLMRobertaXLForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForTokenClassification) (XLMRobertaXLConfig model) - [XLNetConfig](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetConfig) configuration class: [XLNetForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetForTokenClassification) (XLNetConfig model) - [XmodConfig](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodConfig) configuration class: [XmodForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForTokenClassification) (XmodConfig model) - [YosoConfig](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoConfig) configuration class: [YosoForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoForTokenClassification) (YosoConfig model)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

configuration에서 라이브러리의 모델 클래스 중 하나(토큰 분류 헤드를 포함)를 초기화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTokenClassification

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForTokenClassification.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForTokenClassification.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **albert** -- [AlbertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForTokenClassification) (AlbertConfig 모델)
- **apertus** -- [ApertusForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/apertus#transformers.ApertusForTokenClassification) (ApertusConfig 모델)
- **arcee** -- [ArceeForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeForTokenClassification) (ArceeConfig 모델)
- **axk1** -- [AXK1ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/axk1#transformers.AXK1ForTokenClassification) (AXK1Config 모델)
- **axk2** -- [AXK2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/axk2#transformers.AXK2ForTokenClassification) (AXK2Config 모델)
- **bert** -- [BertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForTokenClassification) (BertConfig 모델)
- **big_bird** -- [BigBirdForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForTokenClassification) (BigBirdConfig 모델)
- **biogpt** -- [BioGptForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/biogpt#transformers.BioGptForTokenClassification) (BioGptConfig 모델)
- **bloom** -- [BloomForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomForTokenClassification) (BloomConfig 모델)
- **bros** -- [BrosForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/bros#transformers.BrosForTokenClassification) (BrosConfig 모델)
- **camembert** -- [CamembertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForTokenClassification) (CamembertConfig 모델)
- **canine** -- [CanineForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineForTokenClassification) (CanineConfig 모델)
- **convbert** -- [ConvBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertForTokenClassification) (ConvBertConfig 모델)
- **data2vec-text** -- [Data2VecTextForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForTokenClassification) (Data2VecTextConfig 모델)
- **deberta** -- [DebertaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaForTokenClassification) (DebertaConfig 모델)
- **deberta-v2** -- [DebertaV2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForTokenClassification) (DebertaV2Config 모델)
- **deepseek_v3** -- [DeepseekV3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/deepseek_v3#transformers.DeepseekV3ForTokenClassification) (DeepseekV3Config 모델)
- **diffllama** -- [DiffLlamaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaForTokenClassification) (DiffLlamaConfig 모델)
- **distilbert** -- [DistilBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForTokenClassification) (DistilBertConfig 모델)
- **electra** -- [ElectraForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForTokenClassification) (ElectraConfig 모델)
- **ernie** -- [ErnieForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForTokenClassification) (ErnieConfig 모델)
- **esm** -- [EsmForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/esm#transformers.EsmForTokenClassification) (EsmConfig 모델)
- **esmc** -- [EsmcForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/esmc#transformers.EsmcForTokenClassification) (EsmcConfig 모델)
- **eurobert** -- [EuroBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/eurobert#transformers.EuroBertForTokenClassification) (EuroBertConfig 모델)
- **exaone4** -- [Exaone4ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4ForTokenClassification) (Exaone4Config 모델)
- **falcon** -- [FalconForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconForTokenClassification) (FalconConfig 모델)
- **flaubert** -- [FlaubertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertForTokenClassification) (FlaubertConfig 모델)
- **fnet** -- [FNetForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForTokenClassification) (FNetConfig 모델)
- **funnel** -- [FunnelForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForTokenClassification) (FunnelConfig 모델)
- **gemma** -- [GemmaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma#transformers.GemmaForTokenClassification) (GemmaConfig 모델)
- **gemma2** -- [Gemma2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gemma2#transformers.Gemma2ForTokenClassification) (Gemma2Config 모델)
- **glm** -- [GlmForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/glm#transformers.GlmForTokenClassification) (GlmConfig 모델)
- **glm4** -- [Glm4ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/glm4#transformers.Glm4ForTokenClassification) (Glm4Config 모델)
- **gpt-sw3** -- [GPT2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2ForTokenClassification) (GPT2Config 모델)
- **gpt2** -- [GPT2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2ForTokenClassification) (GPT2Config 모델)
- **gpt_bigcode** -- [GPTBigCodeForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_bigcode#transformers.GPTBigCodeForTokenClassification) (GPTBigCodeConfig 모델)
- **gpt_neo** -- [GPTNeoForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoForTokenClassification) (GPTNeoConfig 모델)
- **gpt_neox** -- [GPTNeoXForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXForTokenClassification) (GPTNeoXConfig 모델)
- **gpt_oss** -- [GptOssForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gpt_oss#transformers.GptOssForTokenClassification) (GptOssConfig 모델)
- **gte** -- [GteForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/gte#transformers.GteForTokenClassification) (GteConfig 모델)
- **helium** -- [HeliumForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/helium#transformers.HeliumForTokenClassification) (HeliumConfig 모델)
- **ibert** -- [IBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForTokenClassification) (IBertConfig 모델)
- **jina_embeddings_v3** -- [JinaEmbeddingsV3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3ForTokenClassification) (JinaEmbeddingsV3Config 모델)
- **layoutlm** -- [LayoutLMForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMForTokenClassification) (LayoutLMConfig 모델)
- **layoutlmv2** -- [LayoutLMv2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2ForTokenClassification) (LayoutLMv2Config 모델)
- **layoutlmv3** -- [LayoutLMv3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3ForTokenClassification) (LayoutLMv3Config 모델)
- **lilt** -- [LiltForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltForTokenClassification) (LiltConfig 모델)
- **llama** -- [LlamaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/llama#transformers.LlamaForTokenClassification) (LlamaConfig 모델)
- **longformer** -- [LongformerForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForTokenClassification) (LongformerConfig 모델)
- **luke** -- [LukeForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForTokenClassification) (LukeConfig 모델)
- **markuplm** -- [MarkupLMForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMForTokenClassification) (MarkupLMConfig 모델)
- **megatron-bert** -- [MegatronBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForTokenClassification) (MegatronBertConfig 모델)
- **minimax** -- [MiniMaxForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxForTokenClassification) (MiniMaxConfig 모델)
- **ministral** -- [MinistralForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralForTokenClassification) (MinistralConfig 모델)
- **ministral3** -- [Ministral3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3ForTokenClassification) (Ministral3Config 모델)
- **mistral** -- [MistralForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralForTokenClassification) (MistralConfig 모델)
- **mistral4** -- [Mistral4ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4ForTokenClassification) (Mistral4Config 모델)
- **mixtral** -- [MixtralForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralForTokenClassification) (MixtralConfig 모델)
- **mobilebert** -- [MobileBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForTokenClassification) (MobileBertConfig 모델)
- **modernbert** -- [ModernBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertForTokenClassification) (ModernBertConfig 모델)
- **modernvbert** -- [ModernVBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/modernvbert#transformers.ModernVBertForTokenClassification) (ModernVBertConfig 모델)
- **mpnet** -- [MPNetForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForTokenClassification) (MPNetConfig 모델)
- **mpt** -- [MptForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptForTokenClassification) (MptConfig 모델)
- **mra** -- [MraForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForTokenClassification) (MraConfig 모델)
- **mt5** -- [MT5ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5ForTokenClassification) (MT5Config 모델)
- **nemotron** -- [NemotronForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronForTokenClassification) (NemotronConfig 모델)
- **nomic_bert** -- [NomicBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/nomic_bert#transformers.NomicBertForTokenClassification) (NomicBertConfig 모델)
- **nystromformer** -- [NystromformerForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerForTokenClassification) (NystromformerConfig 모델)
- **openai_privacy_filter** -- [OpenAIPrivacyFilterForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/openai_privacy_filter#transformers.OpenAIPrivacyFilterForTokenClassification) (OpenAIPrivacyFilterConfig 모델)
- **persimmon** -- [PersimmonForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/persimmon#transformers.PersimmonForTokenClassification) (PersimmonConfig 모델)
- **phi** -- [PhiForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/phi#transformers.PhiForTokenClassification) (PhiConfig 모델)
- **phi3** -- [Phi3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/phi3#transformers.Phi3ForTokenClassification) (Phi3Config 모델)
- **qwen2** -- [Qwen2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2ForTokenClassification) (Qwen2Config 모델)
- **qwen2_moe** -- [Qwen2MoeForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeForTokenClassification) (Qwen2MoeConfig 모델)
- **qwen3** -- [Qwen3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3ForTokenClassification) (Qwen3Config 모델)
- **qwen3_5** -- [Qwen3_5ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForTokenClassification) (Qwen3_5Config 모델)
- **qwen3_asr** -- [Qwen3ASRForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRForTokenClassification) (Qwen3ASRConfig 모델)
- **qwen3_moe** -- [Qwen3MoeForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeForTokenClassification) (Qwen3MoeConfig 모델)
- **qwen3_next** -- [Qwen3NextForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextForTokenClassification) (Qwen3NextConfig 모델)
- **rembert** -- [RemBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForTokenClassification) (RemBertConfig 모델)
- **roberta** -- [RobertaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForTokenClassification) (RobertaConfig 모델)
- **roberta-prelayernorm** -- [RobertaPreLayerNormForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForTokenClassification) (RobertaPreLayerNormConfig 모델)
- **roc_bert** -- [RoCBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForTokenClassification) (RoCBertConfig 모델)
- **roformer** -- [RoFormerForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForTokenClassification) (RoFormerConfig 모델)
- **seed_oss** -- [SeedOssForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssForTokenClassification) (SeedOssConfig 모델)
- **smollm3** -- [SmolLM3ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3ForTokenClassification) (SmolLM3Config 모델)
- **squeezebert** -- [SqueezeBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForTokenClassification) (SqueezeBertConfig 모델)
- **stablelm** -- [StableLmForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/stablelm#transformers.StableLmForTokenClassification) (StableLmConfig 모델)
- **starcoder2** -- [Starcoder2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/starcoder2#transformers.Starcoder2ForTokenClassification) (Starcoder2Config 모델)
- **t5** -- [T5ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5ForTokenClassification) (T5Config 모델)
- **t5gemma** -- [T5GemmaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/t5gemma#transformers.T5GemmaForTokenClassification) (T5GemmaConfig 모델)
- **t5gemma2** -- [T5Gemma2ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForTokenClassification) (T5Gemma2Config 모델)
- **ultrabert** -- [UltrabertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertForTokenClassification) (UltrabertConfig 모델)
- **ultravbert** -- [UltraVBertForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/ultravbert#transformers.UltraVBertForTokenClassification) (UltraVBertConfig 모델)
- **umt5** -- [UMT5ForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5ForTokenClassification) (UMT5Config 모델)
- **xlm** -- [XLMForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMForTokenClassification) (XLMConfig 모델)
- **xlm-roberta** -- [XLMRobertaForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForTokenClassification) (XLMRobertaConfig 모델)
- **xlm-roberta-xl** -- [XLMRobertaXLForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForTokenClassification) (XLMRobertaXLConfig 모델)
- **xlnet** -- [XLNetForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetForTokenClassification) (XLNetConfig 모델)
- **xmod** -- [XmodForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForTokenClassification) (XmodConfig 모델)
- **yoso** -- [YosoForTokenClassification](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoForTokenClassification) (YosoConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTokenClassification

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForTokenClassification.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForTokenClassification.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForQuestionAnswering[[transformers.AutoModelForQuestionAnswering]]

#### transformers.AutoModelForQuestionAnswering[[transformers.AutoModelForQuestionAnswering]]

```python
transformers.AutoModelForQuestionAnswering(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2391)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 라이브러리의 모델 클래스 중 하나(질문 답변 헤드가 포함된)로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForQuestionAnswering.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다: - [AlbertConfig](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertConfig) 구성 클래스: [AlbertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForQuestionAnswering) (AlbertConfig 모델) - [ArceeConfig](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeConfig) 구성 클래스: [ArceeForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeForQuestionAnswering) (ArceeConfig 모델) - [BartConfig](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartConfig) 구성 클래스: [BartForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForQuestionAnswering) (BartConfig 모델) - [BertConfig](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertConfig) 구성 클래스: [BertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForQuestionAnswering) (BertConfig 모델) - [BigBirdConfig](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdConfig) 구성 클래스: [BigBirdForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForQuestionAnswering) (BigBirdConfig 모델) - [BigBirdPegasusConfig](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusConfig) 구성 클래스: [BigBirdPegasusForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusForQuestionAnswering) (BigBirdPegasusConfig 모델) - [BloomConfig](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomConfig) 구성 클래스: [BloomForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomForQuestionAnswering) (BloomConfig 모델) - [CamembertConfig](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertConfig) 구성 클래스: [CamembertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForQuestionAnswering) (CamembertConfig 모델) - [CanineConfig](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineConfig) 구성 클래스: [CanineForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineForQuestionAnswering) (CanineConfig 모델) - [ConvBertConfig](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertConfig) 구성 클래스: [ConvBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertForQuestionAnswering) (ConvBertConfig 모델) - [Data2VecTextConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextConfig) 구성 클래스: [Data2VecTextForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForQuestionAnswering) (Data2VecTextConfig 모델) - [DebertaConfig](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaConfig) 구성 클래스: [DebertaForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaForQuestionAnswering) (DebertaConfig 모델) - [DebertaV2Config](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2Config) 구성 클래스: [DebertaV2ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForQuestionAnswering) (DebertaV2Config 모델) - [DiffLlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaConfig) 구성 클래스: [DiffLlamaForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaForQuestionAnswering) (DiffLlamaConfig 모델) - [DistilBertConfig](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertConfig) 구성 클래스: [DistilBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForQuestionAnswering) (DistilBertConfig 모델) - [ElectraConfig](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraConfig) 구성 클래스: [ElectraForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForQuestionAnswering) (ElectraConfig 모델) - [ErnieConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieConfig) 구성 클래스: [ErnieForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForQuestionAnswering) (ErnieConfig 모델) - [Exaone4Config](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4Config) 구성 클래스: [Exaone4ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4ForQuestionAnswering) (Exaone4Config 모델) - [FNetConfig](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetConfig) 구성 클래스: [FNetForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForQuestionAnswering) (FNetConfig 모델) - [FalconConfig](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconConfig) 구성 클래스: [FalconForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconForQuestionAnswering) (FalconConfig 모델) - [FlaubertConfig](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertConfig) 구성 클래스: [FlaubertForQuestionAnsweringSimple](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertForQuestionAnsweringSimple) (FlaubertConfig 모델) - [FunnelConfig](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelConfig) 구성 클래스: [FunnelForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForQuestionAnswering) (FunnelConfig 모델) - [GPT2Config](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2Config) 구성 클래스: [GPT2ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2ForQuestionAnswering) (GPT2Config 모델) - [GPTJConfig](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJConfig) 구성 클래스: [GPTJForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJForQuestionAnswering) (GPTJConfig 모델) - [GPTNeoConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoConfig) 구성 클래스: [GPTNeoForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoForQuestionAnswering) (GPTNeoConfig 모델) - [GPTNeoXConfig](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXConfig) 구성 클래스: [GPTNeoXForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXForQuestionAnswering) (GPTNeoXConfig 모델) - [IBertConfig](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertConfig) 구성 클래스: [IBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForQuestionAnswering) (IBertConfig 모델) - [JinaEmbeddingsV3Config](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3Config) 구성 클래스: [JinaEmbeddingsV3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3ForQuestionAnswering) (JinaEmbeddingsV3Config 모델) - [LEDConfig](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.LEDConfig) 구성 클래스: [LEDForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.LEDForQuestionAnswering) (LEDConfig 모델) - [LayoutLMv2Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2Config) 구성 클래스: [LayoutLMv2ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2ForQuestionAnswering) (LayoutLMv2Config 모델) - [LayoutLMv3Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3Config) 구성 클래스: [LayoutLMv3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3ForQuestionAnswering) (LayoutLMv3Config 모델) - [LiltConfig](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltConfig) 구성 클래스: [LiltForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltForQuestionAnswering) (LiltConfig 모델) - [LlamaConfig](/docs/transformers/v5.23.0/ko/model_doc/llama2#transformers.LlamaConfig) 구성 클래스: [LlamaForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/llama#transformers.LlamaForQuestionAnswering) (LlamaConfig 모델) - [LongformerConfig](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerConfig) 구성 클래스: [LongformerForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForQuestionAnswering) (LongformerConfig 모델) - [LukeConfig](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeConfig) 구성 클래스: [LukeForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForQuestionAnswering) (LukeConfig 모델) - [LxmertConfig](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.LxmertConfig) 구성 클래스: [LxmertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.LxmertForQuestionAnswering) (LxmertConfig 모델) - [MBartConfig](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartConfig) 구성 클래스: [MBartForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartForQuestionAnswering) (MBartConfig 모델) - [MPNetConfig](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetConfig) 구성 클래스: [MPNetForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForQuestionAnswering) (MPNetConfig 모델) - [MT5Config](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5Config) 구성 클래스: [MT5ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5ForQuestionAnswering) (MT5Config 모델) - [MarkupLMConfig](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMConfig) 구성 클래스: [MarkupLMForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMForQuestionAnswering) (MarkupLMConfig 모델) - [MegatronBertConfig](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertConfig) 구성 클래스: [MegatronBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForQuestionAnswering) (MegatronBertConfig 모델) - [MiniMaxConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxConfig) 구성 클래스: [MiniMaxForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxForQuestionAnswering) (MiniMaxConfig 모델) - [Ministral3Config](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3Config) 구성 클래스: [Ministral3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3ForQuestionAnswering) (Ministral3Config 모델) - [MinistralConfig](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralConfig) 구성 클래스: [MinistralForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralForQuestionAnswering) (MinistralConfig 모델) - [MistralConfig](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralConfig) 구성 클래스: [MistralForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralForQuestionAnswering) (MistralConfig 모델) - [MixtralConfig](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralConfig) 구성 클래스: [MixtralForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralForQuestionAnswering) (MixtralConfig 모델) - [MobileBertConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertConfig) 구성 클래스: [MobileBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForQuestionAnswering) (MobileBertConfig 모델) - [ModernBertConfig](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertConfig) 구성 클래스: [ModernBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertForQuestionAnswering) (ModernBertConfig 모델) - [MptConfig](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptConfig) 구성 클래스: [MptForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptForQuestionAnswering) (MptConfig 모델) - [MraConfig](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraConfig) 구성 클래스: [MraForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForQuestionAnswering) (MraConfig 모델) - [MvpConfig](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpConfig) 구성 클래스: [MvpForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForQuestionAnswering) (MvpConfig 모델) - [NemotronConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronConfig) 구성 클래스: [NemotronForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronForQuestionAnswering) (NemotronConfig 모델) - [NystromformerConfig](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerConfig) 구성 클래스: [NystromformerForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerForQuestionAnswering) (NystromformerConfig 모델) - [OPTConfig](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTConfig) 구성 클래스: [OPTForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTForQuestionAnswering) (OPTConfig 모델) - [Qwen2Config](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Config) 구성 클래스: [Qwen2ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2ForQuestionAnswering) (Qwen2Config 모델) - [Qwen2MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeConfig) 구성 클래스: [Qwen2MoeForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeForQuestionAnswering) (Qwen2MoeConfig 모델) - [Qwen3Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3Config) 구성 클래스: [Qwen3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3ForQuestionAnswering) (Qwen3Config 모델) - [Qwen3MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeConfig) 구성 클래스: [Qwen3MoeForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeForQuestionAnswering) (Qwen3MoeConfig 모델) - [Qwen3NextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextConfig) 구성 클래스: [Qwen3NextForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextForQuestionAnswering) (Qwen3NextConfig 모델) - [ReformerConfig](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerConfig) 구성 클래스: [ReformerForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerForQuestionAnswering) (ReformerConfig 모델) - [RemBertConfig](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertConfig) 구성 클래스: [RemBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForQuestionAnswering) (RemBertConfig 모델) - [RoCBertConfig](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertConfig) 구성 클래스: [RoCBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForQuestionAnswering) (RoCBertConfig 모델) - [RoFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerConfig) 구성 클래스: [RoFormerForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForQuestionAnswering) (RoFormerConfig 모델) - [RobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaConfig) 구성 클래스: [RobertaForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForQuestionAnswering) (RobertaConfig 모델) - [RobertaPreLayerNormConfig](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormConfig) 구성 클래스: [RobertaPreLayerNormForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForQuestionAnswering) (RobertaPreLayerNormConfig 모델) - [SeedOssConfig](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssConfig) 구성 클래스: [SeedOssForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssForQuestionAnswering) (SeedOssConfig 모델) - [SmolLM3Config](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3Config) 구성 클래스: [SmolLM3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3ForQuestionAnswering) (SmolLM3Config 모델) - [SplinterConfig](/docs/transformers/v5.23.0/ko/model_doc/splinter#transformers.SplinterConfig) 구성 클래스: [SplinterForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/splinter#transformers.SplinterForQuestionAnswering) (SplinterConfig 모델) - [SqueezeBertConfig](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertConfig) 구성 클래스: [SqueezeBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForQuestionAnswering) (SqueezeBertConfig 모델) - [T5Config](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5Config) 구성 클래스: [T5ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5ForQuestionAnswering) (T5Config 모델) - [UMT5Config](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5Config) 구성 클래스: [UMT5ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5ForQuestionAnswering) (UMT5Config 모델) - [UltrabertConfig](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertConfig) 구성 클래스: [UltrabertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertForQuestionAnswering) (UltrabertConfig 모델) - [XLMConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMConfig) 구성 클래스: [XLMForQuestionAnsweringSimple](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMForQuestionAnsweringSimple) (XLMConfig 모델) - [XLMRobertaConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaConfig) 구성 클래스: [XLMRobertaForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForQuestionAnswering) (XLMRobertaConfig 모델) - [XLMRobertaXLConfig](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLConfig) 구성 클래스: [XLMRobertaXLForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForQuestionAnswering) (XLMRobertaXLConfig 모델) - [XLNetConfig](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetConfig) 구성 클래스: [XLNetForQuestionAnsweringSimple](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetForQuestionAnsweringSimple) (XLNetConfig 모델) - [XmodConfig](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodConfig) 구성 클래스: [XmodForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForQuestionAnswering) (XmodConfig 모델) - [YosoConfig](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoConfig) 구성 클래스: [YosoForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoForQuestionAnswering) (YosoConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForQuestionAnswering

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForQuestionAnswering.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForQuestionAnswering.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **albert** -- [AlbertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/albert#transformers.AlbertForQuestionAnswering) (AlbertConfig 모델)
- **arcee** -- [ArceeForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/arcee#transformers.ArceeForQuestionAnswering) (ArceeConfig 모델)
- **bart** -- [BartForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/bart#transformers.BartForQuestionAnswering) (BartConfig 모델)
- **bert** -- [BertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/bert#transformers.BertForQuestionAnswering) (BertConfig 모델)
- **big_bird** -- [BigBirdForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/big_bird#transformers.BigBirdForQuestionAnswering) (BigBirdConfig 모델)
- **bigbird_pegasus** -- [BigBirdPegasusForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/bigbird_pegasus#transformers.BigBirdPegasusForQuestionAnswering) (BigBirdPegasusConfig 모델)
- **bloom** -- [BloomForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/bloom#transformers.BloomForQuestionAnswering) (BloomConfig 모델)
- **camembert** -- [CamembertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/camembert#transformers.CamembertForQuestionAnswering) (CamembertConfig 모델)
- **canine** -- [CanineForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/canine#transformers.CanineForQuestionAnswering) (CanineConfig 모델)
- **convbert** -- [ConvBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/convbert#transformers.ConvBertForQuestionAnswering) (ConvBertConfig 모델)
- **data2vec-text** -- [Data2VecTextForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecTextForQuestionAnswering) (Data2VecTextConfig 모델)
- **deberta** -- [DebertaForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/deberta#transformers.DebertaForQuestionAnswering) (DebertaConfig 모델)
- **deberta-v2** -- [DebertaV2ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/deberta-v2#transformers.DebertaV2ForQuestionAnswering) (DebertaV2Config 모델)
- **diffllama** -- [DiffLlamaForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/diffllama#transformers.DiffLlamaForQuestionAnswering) (DiffLlamaConfig 모델)
- **distilbert** -- [DistilBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/distilbert#transformers.DistilBertForQuestionAnswering) (DistilBertConfig 모델)
- **electra** -- [ElectraForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/electra#transformers.ElectraForQuestionAnswering) (ElectraConfig 모델)
- **ernie** -- [ErnieForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/ernie#transformers.ErnieForQuestionAnswering) (ErnieConfig 모델)
- **exaone4** -- [Exaone4ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/exaone4#transformers.Exaone4ForQuestionAnswering) (Exaone4Config 모델)
- **falcon** -- [FalconForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/falcon#transformers.FalconForQuestionAnswering) (FalconConfig 모델)
- **flaubert** -- [FlaubertForQuestionAnsweringSimple](/docs/transformers/v5.23.0/ko/model_doc/flaubert#transformers.FlaubertForQuestionAnsweringSimple) (FlaubertConfig 모델)
- **fnet** -- [FNetForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/fnet#transformers.FNetForQuestionAnswering) (FNetConfig 모델)
- **funnel** -- [FunnelForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/funnel#transformers.FunnelForQuestionAnswering) (FunnelConfig 모델)
- **gpt2** -- [GPT2ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/gpt2#transformers.GPT2ForQuestionAnswering) (GPT2Config 모델)
- **gpt_neo** -- [GPTNeoForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/gpt_neo#transformers.GPTNeoForQuestionAnswering) (GPTNeoConfig 모델)
- **gpt_neox** -- [GPTNeoXForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/gpt_neox#transformers.GPTNeoXForQuestionAnswering) (GPTNeoXConfig 모델)
- **gptj** -- [GPTJForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/gptj#transformers.GPTJForQuestionAnswering) (GPTJConfig 모델)
- **ibert** -- [IBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/ibert#transformers.IBertForQuestionAnswering) (IBertConfig 모델)
- **jina_embeddings_v3** -- [JinaEmbeddingsV3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/jina_embeddings_v3#transformers.JinaEmbeddingsV3ForQuestionAnswering) (JinaEmbeddingsV3Config 모델)
- **layoutlmv2** -- [LayoutLMv2ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2ForQuestionAnswering) (LayoutLMv2Config 모델)
- **layoutlmv3** -- [LayoutLMv3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3ForQuestionAnswering) (LayoutLMv3Config 모델)
- **led** -- [LEDForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/led#transformers.LEDForQuestionAnswering) (LEDConfig 모델)
- **lilt** -- [LiltForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/lilt#transformers.LiltForQuestionAnswering) (LiltConfig 모델)
- **llama** -- [LlamaForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/llama#transformers.LlamaForQuestionAnswering) (LlamaConfig 모델)
- **longformer** -- [LongformerForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/longformer#transformers.LongformerForQuestionAnswering) (LongformerConfig 모델)
- **luke** -- [LukeForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/luke#transformers.LukeForQuestionAnswering) (LukeConfig 모델)
- **lxmert** -- [LxmertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/lxmert#transformers.LxmertForQuestionAnswering) (LxmertConfig 모델)
- **markuplm** -- [MarkupLMForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/markuplm#transformers.MarkupLMForQuestionAnswering) (MarkupLMConfig 모델)
- **mbart** -- [MBartForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mbart#transformers.MBartForQuestionAnswering) (MBartConfig 모델)
- **megatron-bert** -- [MegatronBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/megatron-bert#transformers.MegatronBertForQuestionAnswering) (MegatronBertConfig 모델)
- **minimax** -- [MiniMaxForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/minimax#transformers.MiniMaxForQuestionAnswering) (MiniMaxConfig 모델)
- **ministral** -- [MinistralForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/ministral#transformers.MinistralForQuestionAnswering) (MinistralConfig 모델)
- **ministral3** -- [Ministral3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/ministral3#transformers.Ministral3ForQuestionAnswering) (Ministral3Config 모델)
- **mistral** -- [MistralForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mistral#transformers.MistralForQuestionAnswering) (MistralConfig 모델)
- **mixtral** -- [MixtralForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mixtral#transformers.MixtralForQuestionAnswering) (MixtralConfig 모델)
- **mobilebert** -- [MobileBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mobilebert#transformers.MobileBertForQuestionAnswering) (MobileBertConfig 모델)
- **modernbert** -- [ModernBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/modernbert#transformers.ModernBertForQuestionAnswering) (ModernBertConfig 모델)
- **mpnet** -- [MPNetForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mpnet#transformers.MPNetForQuestionAnswering) (MPNetConfig 모델)
- **mpt** -- [MptForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mpt#transformers.MptForQuestionAnswering) (MptConfig 모델)
- **mra** -- [MraForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mra#transformers.MraForQuestionAnswering) (MraConfig 모델)
- **mt5** -- [MT5ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mt5#transformers.MT5ForQuestionAnswering) (MT5Config 모델)
- **mvp** -- [MvpForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/mvp#transformers.MvpForQuestionAnswering) (MvpConfig 모델)
- **nemotron** -- [NemotronForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/nemotron#transformers.NemotronForQuestionAnswering) (NemotronConfig 모델)
- **nystromformer** -- [NystromformerForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/nystromformer#transformers.NystromformerForQuestionAnswering) (NystromformerConfig 모델)
- **opt** -- [OPTForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/opt#transformers.OPTForQuestionAnswering) (OPTConfig 모델)
- **qwen2** -- [Qwen2ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2ForQuestionAnswering) (Qwen2Config 모델)
- **qwen2_moe** -- [Qwen2MoeForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/qwen2_moe#transformers.Qwen2MoeForQuestionAnswering) (Qwen2MoeConfig 모델)
- **qwen3** -- [Qwen3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/qwen3#transformers.Qwen3ForQuestionAnswering) (Qwen3Config 모델)
- **qwen3_moe** -- [Qwen3MoeForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/qwen3_moe#transformers.Qwen3MoeForQuestionAnswering) (Qwen3MoeConfig 모델)
- **qwen3_next** -- [Qwen3NextForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/qwen3_next#transformers.Qwen3NextForQuestionAnswering) (Qwen3NextConfig 모델)
- **reformer** -- [ReformerForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/reformer#transformers.ReformerForQuestionAnswering) (ReformerConfig 모델)
- **rembert** -- [RemBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/rembert#transformers.RemBertForQuestionAnswering) (RemBertConfig 모델)
- **roberta** -- [RobertaForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/roberta#transformers.RobertaForQuestionAnswering) (RobertaConfig 모델)
- **roberta-prelayernorm** -- [RobertaPreLayerNormForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/roberta-prelayernorm#transformers.RobertaPreLayerNormForQuestionAnswering) (RobertaPreLayerNormConfig 모델)
- **roc_bert** -- [RoCBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/roc_bert#transformers.RoCBertForQuestionAnswering) (RoCBertConfig 모델)
- **roformer** -- [RoFormerForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/roformer#transformers.RoFormerForQuestionAnswering) (RoFormerConfig 모델)
- **seed_oss** -- [SeedOssForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/seed_oss#transformers.SeedOssForQuestionAnswering) (SeedOssConfig 모델)
- **smollm3** -- [SmolLM3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/smollm3#transformers.SmolLM3ForQuestionAnswering) (SmolLM3Config 모델)
- **splinter** -- [SplinterForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/splinter#transformers.SplinterForQuestionAnswering) (SplinterConfig 모델)
- **squeezebert** -- [SqueezeBertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/squeezebert#transformers.SqueezeBertForQuestionAnswering) (SqueezeBertConfig 모델)
- **t5** -- [T5ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/t5#transformers.T5ForQuestionAnswering) (T5Config 모델)
- **ultrabert** -- [UltrabertForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/ultrabert#transformers.UltrabertForQuestionAnswering) (UltrabertConfig 모델)
- **umt5** -- [UMT5ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/umt5#transformers.UMT5ForQuestionAnswering) (UMT5Config 모델)
- **xlm** -- [XLMForQuestionAnsweringSimple](/docs/transformers/v5.23.0/ko/model_doc/xlm#transformers.XLMForQuestionAnsweringSimple) (XLMConfig 모델)
- **xlm-roberta** -- [XLMRobertaForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta#transformers.XLMRobertaForQuestionAnswering) (XLMRobertaConfig 모델)
- **xlm-roberta-xl** -- [XLMRobertaXLForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLForQuestionAnswering) (XLMRobertaXLConfig 모델)
- **xlnet** -- [XLNetForQuestionAnsweringSimple](/docs/transformers/v5.23.0/ko/model_doc/xlnet#transformers.XLNetForQuestionAnsweringSimple) (XLNetConfig 모델)
- **xmod** -- [XmodForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/xmod#transformers.XmodForQuestionAnswering) (XmodConfig 모델)
- **yoso** -- [YosoForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/yoso#transformers.YosoForQuestionAnswering) (YosoConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForQuestionAnswering

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForQuestionAnswering.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForQuestionAnswering.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForTextEncoding[[transformers.AutoModelForTextEncoding]]

#### transformers.AutoModelForTextEncoding[[transformers.AutoModelForTextEncoding]]

```python
transformers.AutoModelForTextEncoding(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2325)

## 컴퓨터 비전

다음과 같은 컴퓨터 비전 작업에 사용할 수 있는 자동 클래스가 있습니다.

### AutoModelForDepthEstimation[[transformers.AutoModelForDepthEstimation]]

#### transformers.AutoModelForDepthEstimation[[transformers.AutoModelForDepthEstimation]]

```python
transformers.AutoModelForDepthEstimation(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2529)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 깊이 추정 헤드를 포함하여 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForDepthEstimation.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다: - [CHMv2Config](/docs/transformers/v5.23.0/ko/model_doc/chmv2#transformers.CHMv2Config) 구성 클래스: [CHMv2ForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/chmv2#transformers.CHMv2ForDepthEstimation) (CHMv2Config 모델) - [DPTConfig](/docs/transformers/v5.23.0/ko/model_doc/dpt#transformers.DPTConfig) 구성 클래스: [DPTForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/dpt#transformers.DPTForDepthEstimation) (DPTConfig 모델) - [DepthAnythingConfig](/docs/transformers/v5.23.0/ko/model_doc/depth_anything#transformers.DepthAnythingConfig) 구성 클래스: [DepthAnythingForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/depth_anything#transformers.DepthAnythingForDepthEstimation) (DepthAnythingConfig 모델) - [DepthProConfig](/docs/transformers/v5.23.0/ko/model_doc/depth_pro#transformers.DepthProConfig) 구성 클래스: [DepthProForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/depth_pro#transformers.DepthProForDepthEstimation) (DepthProConfig 모델) - [GLPNConfig](/docs/transformers/v5.23.0/ko/model_doc/glpn#transformers.GLPNConfig) 구성 클래스: [GLPNForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/glpn#transformers.GLPNForDepthEstimation) (GLPNConfig 모델) - [PromptDepthAnythingConfig](/docs/transformers/v5.23.0/ko/model_doc/prompt_depth_anything#transformers.PromptDepthAnythingConfig) 구성 클래스: [PromptDepthAnythingForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/prompt_depth_anything#transformers.PromptDepthAnythingForDepthEstimation) (PromptDepthAnythingConfig 모델) - [Tipsv2DptConfig](/docs/transformers/v5.23.0/ko/model_doc/tipsv2_dpt#transformers.Tipsv2DptConfig) 구성 클래스: [Tipsv2DptForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/tipsv2_dpt#transformers.Tipsv2DptForDepthEstimation) (Tipsv2DptConfig 모델) - [ZoeDepthConfig](/docs/transformers/v5.23.0/ko/model_doc/zoedepth#transformers.ZoeDepthConfig) 구성 클래스: [ZoeDepthForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/zoedepth#transformers.ZoeDepthForDepthEstimation) (ZoeDepthConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForDepthEstimation

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForDepthEstimation.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForDepthEstimation.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **chmv2** -- [CHMv2ForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/chmv2#transformers.CHMv2ForDepthEstimation) (CHMv2Config 모델)
- **depth_anything** -- [DepthAnythingForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/depth_anything#transformers.DepthAnythingForDepthEstimation) (DepthAnythingConfig 모델)
- **depth_pro** -- [DepthProForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/depth_pro#transformers.DepthProForDepthEstimation) (DepthProConfig 모델)
- **dpt** -- [DPTForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/dpt#transformers.DPTForDepthEstimation) (DPTConfig 모델)
- **glpn** -- [GLPNForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/glpn#transformers.GLPNForDepthEstimation) (GLPNConfig 모델)
- **prompt_depth_anything** -- [PromptDepthAnythingForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/prompt_depth_anything#transformers.PromptDepthAnythingForDepthEstimation) (PromptDepthAnythingConfig 모델)
- **tipsv2_dpt** -- [Tipsv2DptForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/tipsv2_dpt#transformers.Tipsv2DptForDepthEstimation) (Tipsv2DptConfig 모델)
- **zoedepth** -- [ZoeDepthForDepthEstimation](/docs/transformers/v5.23.0/ko/model_doc/zoedepth#transformers.ZoeDepthForDepthEstimation) (ZoeDepthConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForDepthEstimation

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForDepthEstimation.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForDepthEstimation.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForNormalEstimation[[transformers.AutoModelForNormalEstimation]]

#### transformers.AutoModelForNormalEstimation[[transformers.AutoModelForNormalEstimation]]

```python
transformers.AutoModelForNormalEstimation(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2676)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로(일반적인 추정 헤드를 포함하여) 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForNormalEstimation.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:
- [Sapiens2Config](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2Config) 구성 클래스: [Sapiens2ForNormalEstimation](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2ForNormalEstimation) (Sapiens2Config 모델)
- [Tipsv2DptConfig](/docs/transformers/v5.23.0/ko/model_doc/tipsv2_dpt#transformers.Tipsv2DptConfig) 구성 클래스: [Tipsv2DptForNormalEstimation](/docs/transformers/v5.23.0/ko/model_doc/tipsv2_dpt#transformers.Tipsv2DptForNormalEstimation) (Tipsv2DptConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForNormalEstimation

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForNormalEstimation.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForNormalEstimation.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **sapiens2** -- [Sapiens2ForNormalEstimation](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2ForNormalEstimation) (Sapiens2Config 모델)
- **tipsv2_dpt** -- [Tipsv2DptForNormalEstimation](/docs/transformers/v5.23.0/ko/model_doc/tipsv2_dpt#transformers.Tipsv2DptForNormalEstimation) (Tipsv2DptConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForNormalEstimation

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForNormalEstimation.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForNormalEstimation.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForPointmapEstimation[[transformers.AutoModelForPointmapEstimation]]

#### transformers.AutoModelForPointmapEstimation[[transformers.AutoModelForPointmapEstimation]]

```python
transformers.AutoModelForPointmapEstimation(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2683)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다 (pointmap 추정 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForPointmapEstimation.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:  - [Sapiens2Config](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2Config) 구성 클래스: [Sapiens2ForPointmapEstimation](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2ForPointmapEstimation) (Sapiens2Config 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

configuration에서 라이브러리의 모델 클래스 중 하나(점도 추정 헤드를 포함)를 초기화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForPointmapEstimation

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForPointmapEstimation.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForPointmapEstimation.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **sapiens2** -- [Sapiens2ForPointmapEstimation](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2ForPointmapEstimation) (Sapiens2Config 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForPointmapEstimation

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForPointmapEstimation.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForPointmapEstimation.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForImageMatting[[transformers.AutoModelForImageMatting]]

#### transformers.AutoModelForImageMatting[[transformers.AutoModelForImageMatting]]

```python
transformers.AutoModelForImageMatting(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2669)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다 (이미지 매팅 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForImageMatting.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:  - [Sapiens2Config](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2Config) 구성 클래스: [Sapiens2ForImageMatting](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2ForImageMatting) (Sapiens2Config 모델) - [VitMatteConfig](/docs/transformers/v5.23.0/ko/model_doc/vitmatte#transformers.VitMatteConfig) 구성 클래스: [VitMatteForImageMatting](/docs/transformers/v5.23.0/ko/model_doc/vitmatte#transformers.VitMatteForImageMatting) (VitMatteConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForImageMatting

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForImageMatting.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForImageMatting.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **sapiens2** -- [Sapiens2ForImageMatting](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2ForImageMatting) (Sapiens2Config 모델)
- **vitmatte** -- [VitMatteForImageMatting](/docs/transformers/v5.23.0/ko/model_doc/vitmatte#transformers.VitMatteForImageMatting) (VitMatteConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForImageMatting

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForImageMatting.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForImageMatting.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForTextRecognition[[transformers.AutoModelForTextRecognition]]

#### transformers.AutoModelForTextRecognition[[transformers.AutoModelForTextRecognition]]

```python
transformers.AutoModelForTextRecognition(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2536)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 텍스트 인식 헤드를 갖춘 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForTextRecognition.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 설정 클래스에 따라 선택됩니다:
- [PPOCRV5MobileRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_mobile_rec#transformers.PPOCRV5MobileRecConfig) 설정 클래스: [PPOCRV5MobileRecForTextRecognition](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_mobile_rec#transformers.PPOCRV5MobileRecForTextRecognition) (PPOCRV5MobileRecConfig 모델)
- [PPOCRV5ServerRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_server_rec#transformers.PPOCRV5ServerRecConfig) 설정 클래스: [PPOCRV5ServerRecForTextRecognition](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_server_rec#transformers.PPOCRV5ServerRecForTextRecognition) (PPOCRV5ServerRecConfig 모델)
- [PPOCRV6SmallRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_rec#transformers.PPOCRV6SmallRecConfig) 설정 클래스: [PPOCRV6SmallRecForTextRecognition](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_rec#transformers.PPOCRV6SmallRecForTextRecognition) (PPOCRV6SmallRecConfig 모델)
- [PPOCRV6TinyRecConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_tiny_rec#transformers.PPOCRV6TinyRecConfig) 설정 클래스: [PPOCRV6TinyRecForTextRecognition](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_tiny_rec#transformers.PPOCRV6TinyRecForTextRecognition) (PPOCRV6TinyRecConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTextRecognition

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForTextRecognition.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForTextRecognition.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

… 라이브러리의 모델 클래스 중 하나(텍스트 인식 헤드를 포함하여)를 사전 학습된 모델에서 초기화합니다.

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **pp_ocrv5_mobile_rec** -- [PPOCRV5MobileRecForTextRecognition](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_mobile_rec#transformers.PPOCRV5MobileRecForTextRecognition) (PPOCRV5MobileRecConfig 모델)
- **pp_ocrv5_server_rec** -- [PPOCRV5ServerRecForTextRecognition](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_server_rec#transformers.PPOCRV5ServerRecForTextRecognition) (PPOCRV5ServerRecConfig 모델)
- **pp_ocrv6_small_rec** -- [PPOCRV6SmallRecForTextRecognition](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_rec#transformers.PPOCRV6SmallRecForTextRecognition) (PPOCRV6SmallRecConfig 모델)
- **pp_ocrv6_tiny_rec** -- [PPOCRV6TinyRecForTextRecognition](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_tiny_rec#transformers.PPOCRV6TinyRecForTextRecognition) (PPOCRV6TinyRecConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTextRecognition

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForTextRecognition.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForTextRecognition.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForTableRecognition[[transformers.AutoModelForTableRecognition]]

#### transformers.AutoModelForTableRecognition[[transformers.AutoModelForTableRecognition]]

```python
transformers.AutoModelForTableRecognition(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2543)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 테이블 인식 헤드를 갖춘 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForTableRecognition.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 설정 클래스에 따라 선택됩니다:  - [SLANeXtConfig](/docs/transformers/v5.23.0/ko/model_doc/slanext#transformers.SLANeXtConfig) 설정 클래스: [SLANeXtForTableRecognition](/docs/transformers/v5.23.0/ko/model_doc/slanext#transformers.SLANeXtForTableRecognition) (SLANeXtConfig 모델) - [SLANetConfig](/docs/transformers/v5.23.0/ko/model_doc/slanet#transformers.SLANetConfig) 설정 클래스: [SLANetForTableRecognition](/docs/transformers/v5.23.0/ko/model_doc/slanet#transformers.SLANetForTableRecognition) (SLANetConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTableRecognition

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForTableRecognition.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForTableRecognition.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **slanet** -- [SLANetForTableRecognition](/docs/transformers/v5.23.0/ko/model_doc/slanet#transformers.SLANetForTableRecognition) (SLANetConfig 모델)
- **slanext** -- [SLANeXtForTableRecognition](/docs/transformers/v5.23.0/ko/model_doc/slanext#transformers.SLANeXtForTableRecognition) (SLANeXtConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTableRecognition

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForTableRecognition.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForTableRecognition.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForImageClassification[[transformers.AutoModelForImageClassification]]

#### transformers.AutoModelForImageClassification[[transformers.AutoModelForImageClassification]]

```python
transformers.AutoModelForImageClassification(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2454)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드를 통해 생성되면, 라이브러리의 모델 클래스 중 하나로(이미지 분류 헤드를 포함하여) 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForImageClassification.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 설정 클래스에 따라 선택됩니다:  - [BeitConfig](/docs/transformers/v5.23.0/ko/model_doc/beit#transformers.BeitConfig) 설정 클래스: [BeitForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/beit#transformers.BeitForImageClassification) (BeitConfig 모델) - [BitConfig](/docs/transformers/v5.23.0/ko/model_doc/bit#transformers.BitConfig) 설정 클래스: [BitForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/bit#transformers.BitForImageClassification) (BitConfig 모델) - [CLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPConfig) 설정 클래스: [CLIPForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPForImageClassification) (CLIPConfig 모델) - [ConvNextConfig](/docs/transformers/v5.23.0/ko/model_doc/convnext#transformers.ConvNextConfig) 설정 클래스: [ConvNextForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/convnext#transformers.ConvNextForImageClassification) (ConvNextConfig 모델) - [ConvNextV2Config](/docs/transformers/v5.23.0/ko/model_doc/convnextv2#transformers.ConvNextV2Config) 설정 클래스: [ConvNextV2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/convnextv2#transformers.ConvNextV2ForImageClassification) (ConvNextV2Config 모델) - [CvtConfig](/docs/transformers/v5.23.0/ko/model_doc/cvt#transformers.CvtConfig) 설정 클래스: [CvtForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/cvt#transformers.CvtForImageClassification) (CvtConfig 모델) - [Data2VecVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecVisionConfig) 설정 클래스: [Data2VecVisionForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecVisionForImageClassification) (Data2VecVisionConfig 모델) - [DeiTConfig](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTConfig) 설정 클래스: [DeiTForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTForImageClassification) 또는 [DeiTForImageClassificationWithTeacher](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTForImageClassificationWithTeacher) (DeiTConfig 모델) - [DinatConfig](/docs/transformers/v5.23.0/ko/model_doc/dinat#transformers.DinatConfig) 설정 클래스: [DinatForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/dinat#transformers.DinatForImageClassification) (DinatConfig 모델) - [Dinov2Config](/docs/transformers/v5.23.0/ko/model_doc/dinov2#transformers.Dinov2Config) 설정 클래스: [Dinov2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/dinov2#transformers.Dinov2ForImageClassification) (Dinov2Config 모델) - [Dinov2WithRegistersConfig](/docs/transformers/v5.23.0/ko/model_doc/dinov2_with_registers#transformers.Dinov2WithRegistersConfig) 설정 클래스: [Dinov2WithRegistersForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/dinov2_with_registers#transformers.Dinov2WithRegistersForImageClassification) (Dinov2WithRegistersConfig 모델) - [DonutSwinConfig](/docs/transformers/v5.23.0/ko/model_doc/donut#transformers.DonutSwinConfig) 설정 클래스: [DonutSwinForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/donut#transformers.DonutSwinForImageClassification) (DonutSwinConfig 모델) - [EfficientNetConfig](/docs/transformers/v5.23.0/ko/model_doc/efficientnet#transformers.EfficientNetConfig) 설정 클래스: [EfficientNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/efficientnet#transformers.EfficientNetForImageClassification) (EfficientNetConfig 모델) - [FocalNetConfig](/docs/transformers/v5.23.0/ko/model_doc/focalnet#transformers.FocalNetConfig) 설정 클래스: [FocalNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/focalnet#transformers.FocalNetForImageClassification) (FocalNetConfig 모델) - [HGNetV2Config](/docs/transformers/v5.23.0/ko/model_doc/hgnet_v2#transformers.HGNetV2Config) 설정 클래스: [HGNetV2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/hgnet_v2#transformers.HGNetV2ForImageClassification) (HGNetV2Config 모델) - [HieraConfig](/docs/transformers/v5.23.0/ko/model_doc/hiera#transformers.HieraConfig) 설정 클래스: [HieraForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/hiera#transformers.HieraForImageClassification) (HieraConfig 모델) - [IJepaConfig](/docs/transformers/v5.23.0/ko/model_doc/ijepa#transformers.IJepaConfig) 설정 클래스: [IJepaForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/ijepa#transformers.IJepaForImageClassification) (IJepaConfig 모델) - [ImageGPTConfig](/docs/transformers/v5.23.0/ko/model_doc/imagegpt#transformers.ImageGPTConfig) 설정 클래스: [ImageGPTForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/imagegpt#transformers.ImageGPTForImageClassification) (ImageGPTConfig 모델) - [LevitConfig](/docs/transformers/v5.23.0/ko/model_doc/levit#transformers.LevitConfig) 설정 클래스: [LevitForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/levit#transformers.LevitForImageClassification) 또는 [LevitForImageClassificationWithTeacher](/docs/transformers/v5.23.0/ko/model_doc/levit#transformers.LevitForImageClassificationWithTeacher) (LevitConfig 모델) - [MetaClip2Config](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2Config) 설정 클래스: [MetaClip2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2ForImageClassification) (MetaClip2Config 모델) - [MobileNetV1Config](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v1#transformers.MobileNetV1Config) 설정 클래스: [MobileNetV1ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v1#transformers.MobileNetV1ForImageClassification) (MobileNetV1Config 모델) - [MobileNetV2Config](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v2#transformers.MobileNetV2Config) 설정 클래스: [MobileNetV2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v2#transformers.MobileNetV2ForImageClassification) (MobileNetV2Config 모델) - [MobileViTConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilevit#transformers.MobileViTConfig) 설정 클래스: [MobileViTForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilevit#transformers.MobileViTForImageClassification) (MobileViTConfig 모델) - [MobileViTV2Config](/docs/transformers/v5.23.0/ko/model_doc/mobilevitv2#transformers.MobileViTV2Config) 설정 클래스: [MobileViTV2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilevitv2#transformers.MobileViTV2ForImageClassification) (MobileViTV2Config 모델) - [PPLCNetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_lcnet#transformers.PPLCNetConfig) 설정 클래스: [PPLCNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/pp_lcnet#transformers.PPLCNetForImageClassification) (PPLCNetConfig 모델) - [PerceiverConfig](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverConfig) 설정 클래스: [PerceiverForImageClassificationLearned](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverForImageClassificationLearned) 또는 [PerceiverForImageClassificationFourier](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverForImageClassificationFourier) 또는 [PerceiverForImageClassificationConvProcessing](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverForImageClassificationConvProcessing) (PerceiverConfig 모델) - [PoolFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/poolformer#transformers.PoolFormerConfig) 설정 클래스: [PoolFormerForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/poolformer#transformers.PoolFormerForImageClassification) (PoolFormerConfig 모델) - [PvtConfig](/docs/transformers/v5.23.0/ko/model_doc/pvt#transformers.PvtConfig) 설정 클래스: [PvtForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/pvt#transformers.PvtForImageClassification) (PvtConfig 모델) - [PvtV2Config](/docs/transformers/v5.23.0/ko/model_doc/pvt_v2#transformers.PvtV2Config) 설정 클래스: [PvtV2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/pvt_v2#transformers.PvtV2ForImageClassification) (PvtV2Config 모델) - [RegNetConfig](/docs/transformers/v5.23.0/ko/model_doc/regnet#transformers.RegNetConfig) 설정 클래스: [RegNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/regnet#transformers.RegNetForImageClassification) (RegNetConfig 모델) - [ResNetConfig](/docs/transformers/v5.23.0/ko/model_doc/resnet#transformers.ResNetConfig) 설정 클래스: [ResNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/resnet#transformers.ResNetForImageClassification) (ResNetConfig 모델) - [SegformerConfig](/docs/transformers/v5.23.0/ko/model_doc/segformer#transformers.SegformerConfig) 설정 클래스: [SegformerForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/segformer#transformers.SegformerForImageClassification) (SegformerConfig 모델) - [ShieldGemma2Config](/docs/transformers/v5.23.0/ko/model_doc/shieldgemma2#transformers.ShieldGemma2Config) 설정 클래스: [ShieldGemma2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/shieldgemma2#transformers.ShieldGemma2ForImageClassification) (ShieldGemma2Config 모델) - [Siglip2Config](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2Config) 설정 클래스: [Siglip2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2ForImageClassification) (Siglip2Config 모델) - [SiglipConfig](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipConfig) 설정 클래스: [SiglipForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipForImageClassification) (SiglipConfig 모델) - [SwiftFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/swiftformer#transformers.SwiftFormerConfig) 설정 클래스: [SwiftFormerForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/swiftformer#transformers.SwiftFormerForImageClassification) (SwiftFormerConfig 모델) - [SwinConfig](/docs/transformers/v5.23.0/ko/model_doc/swin#transformers.SwinConfig) 설정 클래스: [SwinForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/swin#transformers.SwinForImageClassification) (SwinConfig 모델) - [Swinv2Config](/docs/transformers/v5.23.0/ko/model_doc/swinv2#transformers.Swinv2Config) 설정 클래스: [Swinv2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/swinv2#transformers.Swinv2ForImageClassification) (Swinv2Config 모델) - [TextNetConfig](/docs/transformers/v5.23.0/ko/model_doc/textnet#transformers.TextNetConfig) 설정 클래스: [TextNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/textnet#transformers.TextNetForImageClassification) (TextNetConfig 모델) - [TimmWrapperConfig](/docs/transformers/v5.23.0/ko/model_doc/timm_wrapper#transformers.TimmWrapperConfig) 설정 클래스: [TimmWrapperForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/timm_wrapper#transformers.TimmWrapperForImageClassification) (TimmWrapperConfig 모델) - [ViTConfig](/docs/transformers/v5.23.0/ko/model_doc/vit#transformers.ViTConfig) 설정 클래스: [ViTForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/vit#transformers.ViTForImageClassification) (ViTConfig 모델) - [ViTMSNConfig](/docs/transformers/v5.23.0/ko/model_doc/vit_msn#transformers.ViTMSNConfig) 설정 클래스: [ViTMSNForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/vit_msn#transformers.ViTMSNForImageClassification) (ViTMSNConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForImageClassification

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForImageClassification.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForImageClassification.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **beit** -- [BeitForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/beit#transformers.BeitForImageClassification) (BeitConfig 모델)
- **bit** -- [BitForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/bit#transformers.BitForImageClassification) (BitConfig 모델)
- **clip** -- [CLIPForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPForImageClassification) (CLIPConfig 모델)
- **convnext** -- [ConvNextForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/convnext#transformers.ConvNextForImageClassification) (ConvNextConfig 모델)
- **convnextv2** -- [ConvNextV2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/convnextv2#transformers.ConvNextV2ForImageClassification) (ConvNextV2Config 모델)
- **cvt** -- [CvtForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/cvt#transformers.CvtForImageClassification) (CvtConfig 모델)
- **data2vec-vision** -- [Data2VecVisionForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecVisionForImageClassification) (Data2VecVisionConfig 모델)
- **deit** -- [DeiTForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTForImageClassification) 또는 [DeiTForImageClassificationWithTeacher](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTForImageClassificationWithTeacher) (DeiTConfig 모델)
- **dinat** -- [DinatForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/dinat#transformers.DinatForImageClassification) (DinatConfig 모델)
- **dinov2** -- [Dinov2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/dinov2#transformers.Dinov2ForImageClassification) (Dinov2Config 모델)
- **dinov2_with_registers** -- [Dinov2WithRegistersForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/dinov2_with_registers#transformers.Dinov2WithRegistersForImageClassification) (Dinov2WithRegistersConfig 모델)
- **donut-swin** -- [DonutSwinForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/donut#transformers.DonutSwinForImageClassification) (DonutSwinConfig 모델)
- **efficientnet** -- [EfficientNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/efficientnet#transformers.EfficientNetForImageClassification) (EfficientNetConfig 모델)
- **focalnet** -- [FocalNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/focalnet#transformers.FocalNetForImageClassification) (FocalNetConfig 모델)
- **hgnet_v2** -- [HGNetV2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/hgnet_v2#transformers.HGNetV2ForImageClassification) (HGNetV2Config 모델)
- **hiera** -- [HieraForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/hiera#transformers.HieraForImageClassification) (HieraConfig 모델)
- **ijepa** -- [IJepaForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/ijepa#transformers.IJepaForImageClassification) (IJepaConfig 모델)
- **imagegpt** -- [ImageGPTForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/imagegpt#transformers.ImageGPTForImageClassification) (ImageGPTConfig 모델)
- **levit** -- [LevitForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/levit#transformers.LevitForImageClassification) 또는 [LevitForImageClassificationWithTeacher](/docs/transformers/v5.23.0/ko/model_doc/levit#transformers.LevitForImageClassificationWithTeacher) (LevitConfig 모델)
- **metaclip_2** -- [MetaClip2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2ForImageClassification) (MetaClip2Config 모델)
- **mobilenet_v1** -- [MobileNetV1ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v1#transformers.MobileNetV1ForImageClassification) (MobileNetV1Config 모델)
- **mobilenet_v2** -- [MobileNetV2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v2#transformers.MobileNetV2ForImageClassification) (MobileNetV2Config 모델)
- **mobilevit** -- [MobileViTForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilevit#transformers.MobileViTForImageClassification) (MobileViTConfig 모델)
- **mobilevitv2** -- [MobileViTV2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/mobilevitv2#transformers.MobileViTV2ForImageClassification) (MobileViTV2Config 모델)
- **perceiver** -- [PerceiverForImageClassificationLearned](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverForImageClassificationLearned) 또는 [PerceiverForImageClassificationFourier](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverForImageClassificationFourier) 또는 [PerceiverForImageClassificationConvProcessing](/docs/transformers/v5.23.0/ko/model_doc/perceiver#transformers.PerceiverForImageClassificationConvProcessing) (PerceiverConfig 모델)
- **poolformer** -- [PoolFormerForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/poolformer#transformers.PoolFormerForImageClassification) (PoolFormerConfig 모델)
- **pp_lcnet** -- [PPLCNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/pp_lcnet#transformers.PPLCNetForImageClassification) (PPLCNetConfig 모델)
- **pvt** -- [PvtForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/pvt#transformers.PvtForImageClassification) (PvtConfig 모델)
- **pvt_v2** -- [PvtV2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/pvt_v2#transformers.PvtV2ForImageClassification) (PvtV2Config 모델)
- **regnet** -- [RegNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/regnet#transformers.RegNetForImageClassification) (RegNetConfig 모델)
- **resnet** -- [ResNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/resnet#transformers.ResNetForImageClassification) (ResNetConfig 모델)
- **segformer** -- [SegformerForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/segformer#transformers.SegformerForImageClassification) (SegformerConfig 모델)
- **shieldgemma2** -- [ShieldGemma2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/shieldgemma2#transformers.ShieldGemma2ForImageClassification) (ShieldGemma2Config 모델)
- **siglip** -- [SiglipForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipForImageClassification) (SiglipConfig 모델)
- **siglip2** -- [Siglip2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2ForImageClassification) (Siglip2Config 모델)
- **swiftformer** -- [SwiftFormerForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/swiftformer#transformers.SwiftFormerForImageClassification) (SwiftFormerConfig 모델)
- **swin** -- [SwinForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/swin#transformers.SwinForImageClassification) (SwinConfig 모델)
- **swinv2** -- [Swinv2ForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/swinv2#transformers.Swinv2ForImageClassification) (Swinv2Config 모델)
- **textnet** -- [TextNetForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/textnet#transformers.TextNetForImageClassification) (TextNetConfig 모델)
- **timm_wrapper** -- [TimmWrapperForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/timm_wrapper#transformers.TimmWrapperForImageClassification) (TimmWrapperConfig 모델)
- **vit** -- [ViTForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/vit#transformers.ViTForImageClassification) (ViTConfig 모델)
- **vit_msn** -- [ViTMSNForImageClassification](/docs/transformers/v5.23.0/ko/model_doc/vit_msn#transformers.ViTMSNForImageClassification) (ViTMSNConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForImageClassification

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForImageClassification.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForImageClassification.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForVideoClassification[[transformers.AutoModelForVideoClassification]]

#### transformers.AutoModelForVideoClassification[[transformers.AutoModelForVideoClassification]]

```python
transformers.AutoModelForVideoClassification(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2550)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로(비디오 분류 헤드를 포함하여) 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForVideoClassification.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 설정 클래스에 따라 선택됩니다:  - [TimesformerConfig](/docs/transformers/v5.23.0/ko/model_doc/timesformer#transformers.TimesformerConfig) 설정 클래스: [TimesformerForVideoClassification](/docs/transformers/v5.23.0/ko/model_doc/timesformer#transformers.TimesformerForVideoClassification) (TimesformerConfig 모델) - [VJEPA2Config](/docs/transformers/v5.23.0/ko/model_doc/vjepa2#transformers.VJEPA2Config) 설정 클래스: [VJEPA2ForVideoClassification](/docs/transformers/v5.23.0/ko/model_doc/vjepa2#transformers.VJEPA2ForVideoClassification) (VJEPA2Config 모델) - [VideoMAEConfig](/docs/transformers/v5.23.0/ko/model_doc/videomae#transformers.VideoMAEConfig) 설정 클래스: [VideoMAEForVideoClassification](/docs/transformers/v5.23.0/ko/model_doc/videomae#transformers.VideoMAEForVideoClassification) (VideoMAEConfig 모델) - [VideoPrismVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismVisionConfig) 설정 클래스: [VideoPrismForVideoClassification](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismForVideoClassification) (VideoPrismVisionConfig 모델) - [VivitConfig](/docs/transformers/v5.23.0/ko/model_doc/vivit#transformers.VivitConfig) 설정 클래스: [VivitForVideoClassification](/docs/transformers/v5.23.0/ko/model_doc/vivit#transformers.VivitForVideoClassification) (VivitConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

… 라이브러리의 모델 클래스 중 하나(비디오 분류 헤드를 포함)를 구성에서 인스턴스화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForVideoClassification

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForVideoClassification.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForVideoClassification.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **timesformer** -- [TimesformerForVideoClassification](/docs/transformers/v5.23.0/ko/model_doc/timesformer#transformers.TimesformerForVideoClassification) (TimesformerConfig 모델)
- **videomae** -- [VideoMAEForVideoClassification](/docs/transformers/v5.23.0/ko/model_doc/videomae#transformers.VideoMAEForVideoClassification) (VideoMAEConfig 모델)
- **videoprism_vision_model** -- [VideoPrismForVideoClassification](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismForVideoClassification) (VideoPrismVisionConfig 모델)
- **vivit** -- [VivitForVideoClassification](/docs/transformers/v5.23.0/ko/model_doc/vivit#transformers.VivitForVideoClassification) (VivitConfig 모델)
- **vjepa2** -- [VJEPA2ForVideoClassification](/docs/transformers/v5.23.0/ko/model_doc/vjepa2#transformers.VJEPA2ForVideoClassification) (VJEPA2Config 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForVideoClassification

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForVideoClassification.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForVideoClassification.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForPoseEstimation[[transformers.AutoModelForPoseEstimation]]

#### transformers.AutoModelForPoseEstimation[[transformers.AutoModelForPoseEstimation]]

```python
transformers.AutoModelForPoseEstimation(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2690)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 라이브러리의 모델 클래스 중 하나(포즈 추정 헤드를 포함)로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForPoseEstimation.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:  - [Sapiens2Config](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2Config) 구성 클래스: [Sapiens2ForPoseEstimation](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2ForPoseEstimation) (Sapiens2Config 모델) - [VitPoseConfig](/docs/transformers/v5.23.0/ko/model_doc/vitpose#transformers.VitPoseConfig) 구성 클래스: [VitPoseForPoseEstimation](/docs/transformers/v5.23.0/ko/model_doc/vitpose#transformers.VitPoseForPoseEstimation) (VitPoseConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForPoseEstimation

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForPoseEstimation.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForPoseEstimation.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **sapiens2** -- [Sapiens2ForPoseEstimation](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2ForPoseEstimation) (Sapiens2Config 모델)
- **vitpose** -- [VitPoseForPoseEstimation](/docs/transformers/v5.23.0/ko/model_doc/vitpose#transformers.VitPoseForPoseEstimation) (VitPoseConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForPoseEstimation

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForPoseEstimation.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForPoseEstimation.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForKeypointDetection[[transformers.AutoModelForKeypointDetection]]

#### transformers.AutoModelForKeypointDetection[[transformers.AutoModelForKeypointDetection]]

```python
transformers.AutoModelForKeypointDetection(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2317)

### AutoModelForKeypointMatching[[transformers.AutoModelForKeypointMatching]]

#### transformers.AutoModelForKeypointMatching[[transformers.AutoModelForKeypointMatching]]

```python
transformers.AutoModelForKeypointMatching(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2321)

### AutoModelForMaskedImageModeling[[transformers.AutoModelForMaskedImageModeling]]

#### transformers.AutoModelForMaskedImageModeling[[transformers.AutoModelForMaskedImageModeling]]

```python
transformers.AutoModelForMaskedImageModeling(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2646)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 마스크 이미지 모델링 헤드를 갖춘 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForMaskedImageModeling.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 설정 클래스에 따라 선택됩니다:
- [DeiTConfig](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTConfig) 설정 클래스: [DeiTForMaskedImageModeling](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTForMaskedImageModeling) (DeiTConfig 모델)
- [FocalNetConfig](/docs/transformers/v5.23.0/ko/model_doc/focalnet#transformers.FocalNetConfig) 설정 클래스: [FocalNetForMaskedImageModeling](/docs/transformers/v5.23.0/ko/model_doc/focalnet#transformers.FocalNetForMaskedImageModeling) (FocalNetConfig 모델)
- [SwinConfig](/docs/transformers/v5.23.0/ko/model_doc/swin#transformers.SwinConfig) 설정 클래스: [SwinForMaskedImageModeling](/docs/transformers/v5.23.0/ko/model_doc/swin#transformers.SwinForMaskedImageModeling) (SwinConfig 모델)
- [Swinv2Config](/docs/transformers/v5.23.0/ko/model_doc/swinv2#transformers.Swinv2Config) 설정 클래스: [Swinv2ForMaskedImageModeling](/docs/transformers/v5.23.0/ko/model_doc/swinv2#transformers.Swinv2ForMaskedImageModeling) (Swinv2Config 모델)
- [ViTConfig](/docs/transformers/v5.23.0/ko/model_doc/vit#transformers.ViTConfig) 설정 클래스: [ViTForMaskedImageModeling](/docs/transformers/v5.23.0/ko/model_doc/vit#transformers.ViTForMaskedImageModeling) (ViTConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForMaskedImageModeling

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForMaskedImageModeling.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForMaskedImageModeling.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **deit** -- [DeiTForMaskedImageModeling](/docs/transformers/v5.23.0/ko/model_doc/deit#transformers.DeiTForMaskedImageModeling) (DeiTConfig 모델)
- **focalnet** -- [FocalNetForMaskedImageModeling](/docs/transformers/v5.23.0/ko/model_doc/focalnet#transformers.FocalNetForMaskedImageModeling) (FocalNetConfig 모델)
- **swin** -- [SwinForMaskedImageModeling](/docs/transformers/v5.23.0/ko/model_doc/swin#transformers.SwinForMaskedImageModeling) (SwinConfig 모델)
- **swinv2** -- [Swinv2ForMaskedImageModeling](/docs/transformers/v5.23.0/ko/model_doc/swinv2#transformers.Swinv2ForMaskedImageModeling) (Swinv2Config 모델)
- **vit** -- [ViTForMaskedImageModeling](/docs/transformers/v5.23.0/ko/model_doc/vit#transformers.ViTForMaskedImageModeling) (ViTConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForMaskedImageModeling

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForMaskedImageModeling.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForMaskedImageModeling.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForObjectDetection[[transformers.AutoModelForObjectDetection]]

#### transformers.AutoModelForObjectDetection[[transformers.AutoModelForObjectDetection]]

```python
transformers.AutoModelForObjectDetection(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2513)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 객체 감지 헤드를 갖춘 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForObjectDetection.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:  - [ConditionalDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/conditional_detr#transformers.ConditionalDetrConfig) 구성 클래스: [ConditionalDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/conditional_detr#transformers.ConditionalDetrForObjectDetection) (ConditionalDetrConfig 모델) - [DFineConfig](/docs/transformers/v5.23.0/ko/model_doc/d_fine#transformers.DFineConfig) 구성 클래스: [DFineForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/d_fine#transformers.DFineForObjectDetection) (DFineConfig 모델) - [DabDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/dab-detr#transformers.DabDetrConfig) 구성 클래스: [DabDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/dab-detr#transformers.DabDetrForObjectDetection) (DabDetrConfig 모델) - [DeformableDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/deformable_detr#transformers.DeformableDetrConfig) 구성 클래스: [DeformableDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/deformable_detr#transformers.DeformableDetrForObjectDetection) (DeformableDetrConfig 모델) - [Deimv2Config](/docs/transformers/v5.23.0/ko/model_doc/deimv2#transformers.Deimv2Config) 구성 클래스: [Deimv2ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/deimv2#transformers.Deimv2ForObjectDetection) (Deimv2Config 모델) - [DetrConfig](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrConfig) 구성 클래스: [DetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrForObjectDetection) (DetrConfig 모델) - [LwDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/lw_detr#transformers.LwDetrConfig) 구성 클래스: [LwDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/lw_detr#transformers.LwDetrForObjectDetection) (LwDetrConfig 모델) - [PPDocLayoutV2Config](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v2#transformers.PPDocLayoutV2Config) 구성 클래스: [PPDocLayoutV2ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v2#transformers.PPDocLayoutV2ForObjectDetection) (PPDocLayoutV2Config 모델) - [PPDocLayoutV3Config](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v3#transformers.PPDocLayoutV3Config) 구성 클래스: [PPDocLayoutV3ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v3#transformers.PPDocLayoutV3ForObjectDetection) (PPDocLayoutV3Config 모델) - [PPDocLayoutV4Config](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v4#transformers.PPDocLayoutV4Config) 구성 클래스: [PPDocLayoutV4ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v4#transformers.PPDocLayoutV4ForObjectDetection) (PPDocLayoutV4Config 모델) - [PPOCRV5MobileDetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_mobile_det#transformers.PPOCRV5MobileDetConfig) 구성 클래스: [PPOCRV5MobileDetForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_mobile_det#transformers.PPOCRV5MobileDetForObjectDetection) (PPOCRV5MobileDetConfig 모델) - [PPOCRV5ServerDetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_server_det#transformers.PPOCRV5ServerDetConfig) 구성 클래스: [PPOCRV5ServerDetForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_server_det#transformers.PPOCRV5ServerDetForObjectDetection) (PPOCRV5ServerDetConfig 모델) - [PPOCRV6MediumDetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_medium_det#transformers.PPOCRV6MediumDetConfig) 구성 클래스: [PPOCRV6MediumDetForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_medium_det#transformers.PPOCRV6MediumDetForObjectDetection) (PPOCRV6MediumDetConfig 모델) - [PPOCRV6SmallDetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_det#transformers.PPOCRV6SmallDetConfig) 구성 클래스: [PPOCRV6SmallDetForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_det#transformers.PPOCRV6SmallDetForObjectDetection) (PPOCRV6SmallDetConfig 모델) - [RTDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/rt_detr#transformers.RTDetrConfig) 구성 클래스: [RTDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/rt_detr#transformers.RTDetrForObjectDetection) (RTDetrConfig 모델) - [RTDetrV2Config](/docs/transformers/v5.23.0/ko/model_doc/rt_detr_v2#transformers.RTDetrV2Config) 구성 클래스: [RTDetrV2ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/rt_detr_v2#transformers.RTDetrV2ForObjectDetection) (RTDetrV2Config 모델) - [RfDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/rf_detr#transformers.RfDetrConfig) 구성 클래스: [RfDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/rf_detr#transformers.RfDetrForObjectDetection) (RfDetrConfig 모델) - [TableTransformerConfig](/docs/transformers/v5.23.0/ko/model_doc/table-transformer#transformers.TableTransformerConfig) 구성 클래스: [TableTransformerForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/table-transformer#transformers.TableTransformerForObjectDetection) (TableTransformerConfig 모델) - [YolosConfig](/docs/transformers/v5.23.0/ko/model_doc/yolos#transformers.YolosConfig) 구성 클래스: [YolosForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/yolos#transformers.YolosForObjectDetection) (YolosConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

object detection 헤드가 포함된 라이브러리의 모델 클래스 중 하나를 구성에서 인스턴스화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForObjectDetection

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForObjectDetection.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForObjectDetection.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

object detection head가 포함된 라이브러리의 모델 클래스 중 하나를 사전 학습된 모델에서 초기화합니다.

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **conditional_detr** -- [ConditionalDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/conditional_detr#transformers.ConditionalDetrForObjectDetection) (ConditionalDetrConfig 모델)
- **d_fine** -- [DFineForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/d_fine#transformers.DFineForObjectDetection) (DFineConfig 모델)
- **dab-detr** -- [DabDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/dab-detr#transformers.DabDetrForObjectDetection) (DabDetrConfig 모델)
- **deformable_detr** -- [DeformableDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/deformable_detr#transformers.DeformableDetrForObjectDetection) (DeformableDetrConfig 모델)
- **deimv2** -- [Deimv2ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/deimv2#transformers.Deimv2ForObjectDetection) (Deimv2Config 모델)
- **detr** -- [DetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrForObjectDetection) (DetrConfig 모델)
- **lw_detr** -- [LwDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/lw_detr#transformers.LwDetrForObjectDetection) (LwDetrConfig 모델)
- **pp_doclayout_v2** -- [PPDocLayoutV2ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v2#transformers.PPDocLayoutV2ForObjectDetection) (PPDocLayoutV2Config 모델)
- **pp_doclayout_v3** -- [PPDocLayoutV3ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v3#transformers.PPDocLayoutV3ForObjectDetection) (PPDocLayoutV3Config 모델)
- **pp_doclayout_v4** -- [PPDocLayoutV4ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_doclayout_v4#transformers.PPDocLayoutV4ForObjectDetection) (PPDocLayoutV4Config 모델)
- **pp_ocrv5_mobile_det** -- [PPOCRV5MobileDetForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_mobile_det#transformers.PPOCRV5MobileDetForObjectDetection) (PPOCRV5MobileDetConfig 모델)
- **pp_ocrv5_server_det** -- [PPOCRV5ServerDetForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv5_server_det#transformers.PPOCRV5ServerDetForObjectDetection) (PPOCRV5ServerDetConfig 모델)
- **pp_ocrv6_medium_det** -- [PPOCRV6MediumDetForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_medium_det#transformers.PPOCRV6MediumDetForObjectDetection) (PPOCRV6MediumDetConfig 모델)
- **pp_ocrv6_small_det** -- [PPOCRV6SmallDetForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/pp_ocrv6_small_det#transformers.PPOCRV6SmallDetForObjectDetection) (PPOCRV6SmallDetConfig 모델)
- **rf_detr** -- [RfDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/rf_detr#transformers.RfDetrForObjectDetection) (RfDetrConfig 모델)
- **rt_detr** -- [RTDetrForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/rt_detr#transformers.RTDetrForObjectDetection) (RTDetrConfig 모델)
- **rt_detr_v2** -- [RTDetrV2ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/rt_detr_v2#transformers.RTDetrV2ForObjectDetection) (RTDetrV2Config 모델)
- **table-transformer** -- [TableTransformerForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/table-transformer#transformers.TableTransformerForObjectDetection) (TableTransformerConfig 모델)
- **yolos** -- [YolosForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/yolos#transformers.YolosForObjectDetection) (YolosConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForObjectDetection

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForObjectDetection.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForObjectDetection.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForImageSegmentation[[transformers.AutoModelForImageSegmentation]]

#### transformers.AutoModelForImageSegmentation[[transformers.AutoModelForImageSegmentation]]

```python
transformers.AutoModelForImageSegmentation(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2470)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 라이브러리의 모델 클래스 중 하나로(이미지 분할 헤드를 포함하여) 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForImageSegmentation.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:  - [DetrConfig](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrConfig) 구성 클래스: [DetrForSegmentation](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrForSegmentation) (DetrConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForImageSegmentation

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForImageSegmentation.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForImageSegmentation.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **detr** -- [DetrForSegmentation](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrForSegmentation) (DetrConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForImageSegmentation

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForImageSegmentation.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForImageSegmentation.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForImageToImage[[transformers.AutoModelForImageToImage]]

#### transformers.AutoModelForImageToImage[[transformers.AutoModelForImageToImage]]

```python
transformers.AutoModelForImageToImage(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2329)

### AutoModelForSemanticSegmentation[[transformers.AutoModelForSemanticSegmentation]]

#### transformers.AutoModelForSemanticSegmentation[[transformers.AutoModelForSemanticSegmentation]]

```python
transformers.AutoModelForSemanticSegmentation(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2477)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 시맨틱 세그멘테이션 헤드를 갖춘 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForSemanticSegmentation.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:  - [BeitConfig](/docs/transformers/v5.23.0/ko/model_doc/beit#transformers.BeitConfig) 구성 클래스: [BeitForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/beit#transformers.BeitForSemanticSegmentation) (BeitConfig 모델) - [DPTConfig](/docs/transformers/v5.23.0/ko/model_doc/dpt#transformers.DPTConfig) 구성 클래스: [DPTForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/dpt#transformers.DPTForSemanticSegmentation) (DPTConfig 모델) - [Data2VecVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecVisionConfig) 구성 클래스: [Data2VecVisionForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecVisionForSemanticSegmentation) (Data2VecVisionConfig 모델) - [MobileNetV2Config](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v2#transformers.MobileNetV2Config) 구성 클래스: [MobileNetV2ForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v2#transformers.MobileNetV2ForSemanticSegmentation) (MobileNetV2Config 모델) - [MobileViTConfig](/docs/transformers/v5.23.0/ko/model_doc/mobilevit#transformers.MobileViTConfig) 구성 클래스: [MobileViTForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/mobilevit#transformers.MobileViTForSemanticSegmentation) (MobileViTConfig 모델) - [MobileViTV2Config](/docs/transformers/v5.23.0/ko/model_doc/mobilevitv2#transformers.MobileViTV2Config) 구성 클래스: [MobileViTV2ForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/mobilevitv2#transformers.MobileViTV2ForSemanticSegmentation) (MobileViTV2Config 모델) - [Sapiens2Config](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2Config) 구성 클래스: [Sapiens2ForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2ForSemanticSegmentation) (Sapiens2Config 모델) - [SegformerConfig](/docs/transformers/v5.23.0/ko/model_doc/segformer#transformers.SegformerConfig) 구성 클래스: [SegformerForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/segformer#transformers.SegformerForSemanticSegmentation) (SegformerConfig 모델) - [Tipsv2DptConfig](/docs/transformers/v5.23.0/ko/model_doc/tipsv2_dpt#transformers.Tipsv2DptConfig) 구성 클래스: [Tipsv2DptForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/tipsv2_dpt#transformers.Tipsv2DptForSemanticSegmentation) (Tipsv2DptConfig 모델) - [UperNetConfig](/docs/transformers/v5.23.0/ko/model_doc/upernet#transformers.UperNetConfig) 구성 클래스: [UperNetForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/upernet#transformers.UperNetForSemanticSegmentation) (UperNetConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

… 라이브러리의 모델 클래스 중 하나(시맨틱 세그멘테이션 헤드를 포함)를 구성에서 인스턴스화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForSemanticSegmentation

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForSemanticSegmentation.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForSemanticSegmentation.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **beit** -- [BeitForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/beit#transformers.BeitForSemanticSegmentation) (BeitConfig 모델)
- **data2vec-vision** -- [Data2VecVisionForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecVisionForSemanticSegmentation) (Data2VecVisionConfig 모델)
- **dpt** -- [DPTForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/dpt#transformers.DPTForSemanticSegmentation) (DPTConfig 모델)
- **mobilenet_v2** -- [MobileNetV2ForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/mobilenet_v2#transformers.MobileNetV2ForSemanticSegmentation) (MobileNetV2Config 모델)
- **mobilevit** -- [MobileViTForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/mobilevit#transformers.MobileViTForSemanticSegmentation) (MobileViTConfig 모델)
- **mobilevitv2** -- [MobileViTV2ForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/mobilevitv2#transformers.MobileViTV2ForSemanticSegmentation) (MobileViTV2Config 모델)
- **sapiens2** -- [Sapiens2ForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/sapiens2#transformers.Sapiens2ForSemanticSegmentation) (Sapiens2Config 모델)
- **segformer** -- [SegformerForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/segformer#transformers.SegformerForSemanticSegmentation) (SegformerConfig 모델)
- **tipsv2_dpt** -- [Tipsv2DptForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/tipsv2_dpt#transformers.Tipsv2DptForSemanticSegmentation) (Tipsv2DptConfig 모델)
- **upernet** -- [UperNetForSemanticSegmentation](/docs/transformers/v5.23.0/ko/model_doc/upernet#transformers.UperNetForSemanticSegmentation) (UperNetConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForSemanticSegmentation

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForSemanticSegmentation.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForSemanticSegmentation.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForInstanceSegmentation[[transformers.AutoModelForInstanceSegmentation]]

#### transformers.AutoModelForInstanceSegmentation[[transformers.AutoModelForInstanceSegmentation]]

```python
transformers.AutoModelForInstanceSegmentation(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2504)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 라이브러리의 모델 클래스 중 하나(인스턴스 세그멘테이션 헤드를 포함하여)로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForInstanceSegmentation.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:  - [MaskFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/maskformer#transformers.MaskFormerConfig) 구성 클래스: [MaskFormerForInstanceSegmentation](/docs/transformers/v5.23.0/ko/model_doc/maskformer#transformers.MaskFormerForInstanceSegmentation) (MaskFormerConfig 모델) - [RfDetrConfig](/docs/transformers/v5.23.0/ko/model_doc/rf_detr#transformers.RfDetrConfig) 구성 클래스: [RfDetrForInstanceSegmentation](/docs/transformers/v5.23.0/ko/model_doc/rf_detr#transformers.RfDetrForInstanceSegmentation) (RfDetrConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

configuration에서 라이브러리의 모델 클래스 중 하나(인스턴스 분할 헤드를 포함)를 초기화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForInstanceSegmentation

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForInstanceSegmentation.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForInstanceSegmentation.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **maskformer** -- [MaskFormerForInstanceSegmentation](/docs/transformers/v5.23.0/ko/model_doc/maskformer#transformers.MaskFormerForInstanceSegmentation) (MaskFormerConfig 모델)
- **rf_detr** -- [RfDetrForInstanceSegmentation](/docs/transformers/v5.23.0/ko/model_doc/rf_detr#transformers.RfDetrForInstanceSegmentation) (RfDetrConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForInstanceSegmentation

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForInstanceSegmentation.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForInstanceSegmentation.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForUniversalSegmentation[[transformers.AutoModelForUniversalSegmentation]]

#### transformers.AutoModelForUniversalSegmentation[[transformers.AutoModelForUniversalSegmentation]]

```python
transformers.AutoModelForUniversalSegmentation(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2495)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 범용 모델 클래스입니다 (범용 이미지 분할 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForUniversalSegmentation.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다: - [DetrConfig](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrConfig) 구성 클래스: [DetrForSegmentation](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrForSegmentation) (DetrConfig 모델) - [EomtConfig](/docs/transformers/v5.23.0/ko/model_doc/eomt#transformers.EomtConfig) 구성 클래스: [EomtForUniversalSegmentation](/docs/transformers/v5.23.0/ko/model_doc/eomt#transformers.EomtForUniversalSegmentation) (EomtConfig 모델) - [EomtDinov3Config](/docs/transformers/v5.23.0/ko/model_doc/eomt_dinov3#transformers.EomtDinov3Config) 구성 클래스: [EomtDinov3ForUniversalSegmentation](/docs/transformers/v5.23.0/ko/model_doc/eomt_dinov3#transformers.EomtDinov3ForUniversalSegmentation) (EomtDinov3Config 모델) - [Mask2FormerConfig](/docs/transformers/v5.23.0/ko/model_doc/mask2former#transformers.Mask2FormerConfig) 구성 클래스: [Mask2FormerForUniversalSegmentation](/docs/transformers/v5.23.0/ko/model_doc/mask2former#transformers.Mask2FormerForUniversalSegmentation) (Mask2FormerConfig 모델) - [MaskFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/maskformer#transformers.MaskFormerConfig) 구성 클래스: [MaskFormerForInstanceSegmentation](/docs/transformers/v5.23.0/ko/model_doc/maskformer#transformers.MaskFormerForInstanceSegmentation) (MaskFormerConfig 모델) - [OneFormerConfig](/docs/transformers/v5.23.0/ko/model_doc/oneformer#transformers.OneFormerConfig) 구성 클래스: [OneFormerForUniversalSegmentation](/docs/transformers/v5.23.0/ko/model_doc/oneformer#transformers.OneFormerForUniversalSegmentation) (OneFormerConfig 모델) - [VideomtConfig](/docs/transformers/v5.23.0/ko/model_doc/videomt#transformers.VideomtConfig) 구성 클래스: [VideomtForUniversalSegmentation](/docs/transformers/v5.23.0/ko/model_doc/videomt#transformers.VideomtForUniversalSegmentation) (VideomtConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForUniversalSegmentation

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForUniversalSegmentation.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForUniversalSegmentation.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **detr** -- [DetrForSegmentation](/docs/transformers/v5.23.0/ko/model_doc/detr#transformers.DetrForSegmentation) (DetrConfig 모델)
- **eomt** -- [EomtForUniversalSegmentation](/docs/transformers/v5.23.0/ko/model_doc/eomt#transformers.EomtForUniversalSegmentation) (EomtConfig 모델)
- **eomt_dinov3** -- [EomtDinov3ForUniversalSegmentation](/docs/transformers/v5.23.0/ko/model_doc/eomt_dinov3#transformers.EomtDinov3ForUniversalSegmentation) (EomtDinov3Config 모델)
- **mask2former** -- [Mask2FormerForUniversalSegmentation](/docs/transformers/v5.23.0/ko/model_doc/mask2former#transformers.Mask2FormerForUniversalSegmentation) (Mask2FormerConfig 모델)
- **maskformer** -- [MaskFormerForInstanceSegmentation](/docs/transformers/v5.23.0/ko/model_doc/maskformer#transformers.MaskFormerForInstanceSegmentation) (MaskFormerConfig 모델)
- **oneformer** -- [OneFormerForUniversalSegmentation](/docs/transformers/v5.23.0/ko/model_doc/oneformer#transformers.OneFormerForUniversalSegmentation) (OneFormerConfig 모델)
- **videomt** -- [VideomtForUniversalSegmentation](/docs/transformers/v5.23.0/ko/model_doc/videomt#transformers.VideomtForUniversalSegmentation) (VideomtConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForUniversalSegmentation

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForUniversalSegmentation.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForUniversalSegmentation.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForZeroShotImageClassification[[transformers.AutoModelForZeroShotImageClassification]]

#### transformers.AutoModelForZeroShotImageClassification[[transformers.AutoModelForZeroShotImageClassification]]

```python
transformers.AutoModelForZeroShotImageClassification(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2461)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다 (제로샷 이미지 분류 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForZeroShotImageClassification.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:  - [AlignConfig](/docs/transformers/v5.23.0/ko/model_doc/align#transformers.AlignConfig) 구성 클래스: [AlignModel](/docs/transformers/v5.23.0/ko/model_doc/align#transformers.AlignModel) (AlignConfig 모델) - [AltCLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/altclip#transformers.AltCLIPConfig) 구성 클래스: [AltCLIPModel](/docs/transformers/v5.23.0/ko/model_doc/altclip#transformers.AltCLIPModel) (AltCLIPConfig 모델) - [Blip2Config](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2Config) 구성 클래스: [Blip2ForImageTextRetrieval](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2ForImageTextRetrieval) (Blip2Config 모델) - [BlipConfig](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipConfig) 구성 클래스: [BlipModel](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipModel) (BlipConfig 모델) - [CLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPConfig) 구성 클래스: [CLIPModel](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPModel) (CLIPConfig 모델) - [CLIPSegConfig](/docs/transformers/v5.23.0/ko/model_doc/clipseg#transformers.CLIPSegConfig) 구성 클래스: [CLIPSegModel](/docs/transformers/v5.23.0/ko/model_doc/clipseg#transformers.CLIPSegModel) (CLIPSegConfig 모델) - [ChineseCLIPConfig](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPConfig) 구성 클래스: [ChineseCLIPModel](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPModel) (ChineseCLIPConfig 모델) - [MetaClip2Config](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2Config) 구성 클래스: [MetaClip2Model](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2Model) (MetaClip2Config 모델) - [Siglip2Config](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2Config) 구성 클래스: [Siglip2Model](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2Model) (Siglip2Config 모델) - [SiglipConfig](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipConfig) 구성 클래스: [SiglipModel](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipModel) (SiglipConfig 모델) - [Tipsv2Config](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2Config) 구성 클래스: [Tipsv2Model](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2Model) (Tipsv2Config 모델) - [VideoPrismConfig](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismConfig) 구성 클래스: [VideoPrismClipModel](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismClipModel) (VideoPrismConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForZeroShotImageClassification

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForZeroShotImageClassification.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForZeroShotImageClassification.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **align** -- [AlignModel](/docs/transformers/v5.23.0/ko/model_doc/align#transformers.AlignModel) (AlignConfig 모델)
- **altclip** -- [AltCLIPModel](/docs/transformers/v5.23.0/ko/model_doc/altclip#transformers.AltCLIPModel) (AltCLIPConfig 모델)
- **blip** -- [BlipModel](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipModel) (BlipConfig 모델)
- **blip-2** -- [Blip2ForImageTextRetrieval](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2ForImageTextRetrieval) (Blip2Config 모델)
- **chinese_clip** -- [ChineseCLIPModel](/docs/transformers/v5.23.0/ko/model_doc/chinese_clip#transformers.ChineseCLIPModel) (ChineseCLIPConfig 모델)
- **clip** -- [CLIPModel](/docs/transformers/v5.23.0/ko/model_doc/clip#transformers.CLIPModel) (CLIPConfig 모델)
- **clipseg** -- [CLIPSegModel](/docs/transformers/v5.23.0/ko/model_doc/clipseg#transformers.CLIPSegModel) (CLIPSegConfig 모델)
- **metaclip_2** -- [MetaClip2Model](/docs/transformers/v5.23.0/ko/model_doc/metaclip_2#transformers.MetaClip2Model) (MetaClip2Config 모델)
- **siglip** -- [SiglipModel](/docs/transformers/v5.23.0/ko/model_doc/siglip#transformers.SiglipModel) (SiglipConfig 모델)
- **siglip2** -- [Siglip2Model](/docs/transformers/v5.23.0/ko/model_doc/siglip2#transformers.Siglip2Model) (Siglip2Config 모델)
- **tipsv2** -- [Tipsv2Model](/docs/transformers/v5.23.0/ko/model_doc/tipsv2#transformers.Tipsv2Model) (Tipsv2Config 모델)
- **videoprism** -- [VideoPrismClipModel](/docs/transformers/v5.23.0/ko/model_doc/videoprism#transformers.VideoPrismClipModel) (VideoPrismConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForZeroShotImageClassification

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForZeroShotImageClassification.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForZeroShotImageClassification.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForZeroShotObjectDetection[[transformers.AutoModelForZeroShotObjectDetection]]

#### transformers.AutoModelForZeroShotObjectDetection[[transformers.AutoModelForZeroShotObjectDetection]]

```python
transformers.AutoModelForZeroShotObjectDetection(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2520)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 범용 모델 클래스입니다(제로샷 객체 감지 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForZeroShotObjectDetection.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:  - [GroundingDinoConfig](/docs/transformers/v5.23.0/ko/model_doc/grounding-dino#transformers.GroundingDinoConfig) 구성 클래스: [GroundingDinoForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/grounding-dino#transformers.GroundingDinoForObjectDetection) (GroundingDinoConfig 모델) - [MMGroundingDinoConfig](/docs/transformers/v5.23.0/ko/model_doc/mm-grounding-dino#transformers.MMGroundingDinoConfig) 구성 클래스: [MMGroundingDinoForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/mm-grounding-dino#transformers.MMGroundingDinoForObjectDetection) (MMGroundingDinoConfig 모델) - [OmDetTurboConfig](/docs/transformers/v5.23.0/ko/model_doc/omdet-turbo#transformers.OmDetTurboConfig) 구성 클래스: [OmDetTurboForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/omdet-turbo#transformers.OmDetTurboForObjectDetection) (OmDetTurboConfig 모델) - [OwlViTConfig](/docs/transformers/v5.23.0/ko/model_doc/owlvit#transformers.OwlViTConfig) 구성 클래스: [OwlViTForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/owlvit#transformers.OwlViTForObjectDetection) (OwlViTConfig 모델) - [Owlv2Config](/docs/transformers/v5.23.0/ko/model_doc/owlv2#transformers.Owlv2Config) 구성 클래스: [Owlv2ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/owlv2#transformers.Owlv2ForObjectDetection) (Owlv2Config 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForZeroShotObjectDetection

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForZeroShotObjectDetection.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForZeroShotObjectDetection.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **grounding-dino** -- [GroundingDinoForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/grounding-dino#transformers.GroundingDinoForObjectDetection) (GroundingDinoConfig 모델)
- **mm-grounding-dino** -- [MMGroundingDinoForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/mm-grounding-dino#transformers.MMGroundingDinoForObjectDetection) (MMGroundingDinoConfig 모델)
- **omdet-turbo** -- [OmDetTurboForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/omdet-turbo#transformers.OmDetTurboForObjectDetection) (OmDetTurboConfig 모델)
- **owlv2** -- [Owlv2ForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/owlv2#transformers.Owlv2ForObjectDetection) (Owlv2Config 모델)
- **owlvit** -- [OwlViTForObjectDetection](/docs/transformers/v5.23.0/ko/model_doc/owlvit#transformers.OwlViTForObjectDetection) (OwlViTConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForZeroShotObjectDetection

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForZeroShotObjectDetection.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForZeroShotObjectDetection.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

## 오디오

다음과 같은 오디오 작업에 사용할 수 있는 자동 클래스가 있습니다.

### AutoModelForAudioClassification[[transformers.AutoModelForAudioClassification]]

#### transformers.AutoModelForAudioClassification[[transformers.AutoModelForAudioClassification]]

```python
transformers.AutoModelForAudioClassification(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2581)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 라이브러리의 모델 클래스 중 하나로(오디오 분류 헤드를 포함하여) 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForAudioClassification.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 설정 클래스에 따라 선택됩니다:
- [ASTConfig](/docs/transformers/v5.23.0/ko/model_doc/audio-spectrogram-transformer#transformers.ASTConfig) 설정 클래스: [ASTForAudioClassification](/docs/transformers/v5.23.0/ko/model_doc/audio-spectrogram-transformer#transformers.ASTForAudioClassification) (ASTConfig 모델)
- [Data2VecAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioConfig) 설정 클래스: [Data2VecAudioForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioForSequenceClassification) (Data2VecAudioConfig 모델)
- [HubertConfig](/docs/transformers/v5.23.0/ko/model_doc/hubert#transformers.HubertConfig) 설정 클래스: [HubertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/hubert#transformers.HubertForSequenceClassification) (HubertConfig 모델)
- [SEWConfig](/docs/transformers/v5.23.0/ko/model_doc/sew#transformers.SEWConfig) 설정 클래스: [SEWForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/sew#transformers.SEWForSequenceClassification) (SEWConfig 모델)
- [SEWDConfig](/docs/transformers/v5.23.0/ko/model_doc/sew-d#transformers.SEWDConfig) 설정 클래스: [SEWDForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/sew-d#transformers.SEWDForSequenceClassification) (SEWDConfig 모델)
- [UniSpeechConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechConfig) 설정 클래스: [UniSpeechForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechForSequenceClassification) (UniSpeechConfig 모델)
- [UniSpeechSatConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatConfig) 설정 클래스: [UniSpeechSatForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatForSequenceClassification) (UniSpeechSatConfig 모델)
- [Wav2Vec2BertConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertConfig) 설정 클래스: [Wav2Vec2BertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertForSequenceClassification) (Wav2Vec2BertConfig 모델)
- [Wav2Vec2Config](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Config) 설정 클래스: [Wav2Vec2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2ForSequenceClassification) (Wav2Vec2Config 모델)
- [Wav2Vec2ConformerConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerConfig) 설정 클래스: [Wav2Vec2ConformerForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerForSequenceClassification) (Wav2Vec2ConformerConfig 모델)
- [WavLMConfig](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMConfig) 설정 클래스: [WavLMForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMForSequenceClassification) (WavLMConfig 모델)
- [WhisperConfig](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperConfig) 설정 클래스: [WhisperForAudioClassification](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperForAudioClassification) (WhisperConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

configuration에서 라이브러리의 모델 클래스 중 하나(오디오 분류 헤드를 포함)를 초기화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForAudioClassification

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForAudioClassification.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForAudioClassification.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **audio-spectrogram-transformer** -- [ASTForAudioClassification](/docs/transformers/v5.23.0/ko/model_doc/audio-spectrogram-transformer#transformers.ASTForAudioClassification) (ASTConfig 모델)
- **data2vec-audio** -- [Data2VecAudioForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioForSequenceClassification) (Data2VecAudioConfig 모델)
- **hubert** -- [HubertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/hubert#transformers.HubertForSequenceClassification) (HubertConfig 모델)
- **sew** -- [SEWForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/sew#transformers.SEWForSequenceClassification) (SEWConfig 모델)
- **sew-d** -- [SEWDForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/sew-d#transformers.SEWDForSequenceClassification) (SEWDConfig 모델)
- **unispeech** -- [UniSpeechForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechForSequenceClassification) (UniSpeechConfig 모델)
- **unispeech-sat** -- [UniSpeechSatForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatForSequenceClassification) (UniSpeechSatConfig 모델)
- **wav2vec2** -- [Wav2Vec2ForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2ForSequenceClassification) (Wav2Vec2Config 모델)
- **wav2vec2-bert** -- [Wav2Vec2BertForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertForSequenceClassification) (Wav2Vec2BertConfig 모델)
- **wav2vec2-conformer** -- [Wav2Vec2ConformerForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerForSequenceClassification) (Wav2Vec2ConformerConfig 모델)
- **wavlm** -- [WavLMForSequenceClassification](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMForSequenceClassification) (WavLMConfig 모델)
- **whisper** -- [WhisperForAudioClassification](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperForAudioClassification) (WhisperConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForAudioClassification

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForAudioClassification.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForAudioClassification.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForAudioFrameClassification[[transformers.AutoModelForAudioFrameClassification]]

#### transformers.AutoModelForAudioFrameClassification[[transformers.AutoModelForAudioFrameClassification]]

```python
transformers.AutoModelForAudioFrameClassification(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2618)

이는 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 범용 모델 클래스입니다(오디오 프레임(토큰) 분류 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForAudioFrameClassification.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:  - [Data2VecAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioConfig) 구성 클래스: [Data2VecAudioForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioForAudioFrameClassification) (Data2VecAudioConfig 모델) - [Nemotron3DiarizationConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationConfig) 구성 클래스: [Nemotron3DiarizationForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationForAudioFrameClassification) (Nemotron3DiarizationConfig 모델) - [UniSpeechSatConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatConfig) 구성 클래스: [UniSpeechSatForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatForAudioFrameClassification) (UniSpeechSatConfig 모델) - [Wav2Vec2BertConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertConfig) 구성 클래스: [Wav2Vec2BertForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertForAudioFrameClassification) (Wav2Vec2BertConfig 모델) - [Wav2Vec2Config](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Config) 구성 클래스: [Wav2Vec2ForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2ForAudioFrameClassification) (Wav2Vec2Config 모델) - [Wav2Vec2ConformerConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerConfig) 구성 클래스: [Wav2Vec2ConformerForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerForAudioFrameClassification) (Wav2Vec2ConformerConfig 모델) - [WavLMConfig](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMConfig) 구성 클래스: [WavLMForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMForAudioFrameClassification) (WavLMConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForAudioFrameClassification

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForAudioFrameClassification.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForAudioFrameClassification.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

… 라이브러리의 모델 클래스 중 하나(오디오 프레임(토큰) 분류 헤드를 가진)를 사전 학습된 모델에서 초기화합니다.

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **data2vec-audio** -- [Data2VecAudioForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioForAudioFrameClassification) (Data2VecAudioConfig 모델)
- **nemotron3_diarization** -- [Nemotron3DiarizationForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_diarization#transformers.Nemotron3DiarizationForAudioFrameClassification) (Nemotron3DiarizationConfig 모델)
- **unispeech-sat** -- [UniSpeechSatForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatForAudioFrameClassification) (UniSpeechSatConfig 모델)
- **wav2vec2** -- [Wav2Vec2ForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2ForAudioFrameClassification) (Wav2Vec2Config 모델)
- **wav2vec2-bert** -- [Wav2Vec2BertForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertForAudioFrameClassification) (Wav2Vec2BertConfig 모델)
- **wav2vec2-conformer** -- [Wav2Vec2ConformerForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerForAudioFrameClassification) (Wav2Vec2ConformerConfig 모델)
- **wavlm** -- [WavLMForAudioFrameClassification](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMForAudioFrameClassification) (WavLMConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForAudioFrameClassification

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForAudioFrameClassification.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForAudioFrameClassification.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForCTC[[transformers.AutoModelForCTC]]

#### transformers.AutoModelForCTC[[transformers.AutoModelForCTC]]

```python
transformers.AutoModelForCTC(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2588)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 범용 모델 클래스입니다(연결주의 시간 분류 헤드 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForCTC.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:
- [Data2VecAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioConfig) 구성 클래스: [Data2VecAudioForCTC](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioForCTC) (Data2VecAudioConfig 모델)
- [GraniteSpeech5CTCConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5CTCConfig) 구성 클래스: [GraniteSpeech5ForCTC](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5ForCTC) (GraniteSpeech5CTCConfig 모델)
- [HubertConfig](/docs/transformers/v5.23.0/ko/model_doc/hubert#transformers.HubertConfig) 구성 클래스: [HubertForCTC](/docs/transformers/v5.23.0/ko/model_doc/hubert#transformers.HubertForCTC) (HubertConfig 모델)
- [LasrCTCConfig](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrCTCConfig) 구성 클래스: [LasrForCTC](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrForCTC) (LasrCTCConfig 모델)
- [ParakeetCTCConfig](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetCTCConfig) 구성 클래스: [ParakeetForCTC](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForCTC) (ParakeetCTCConfig 모델)
- [SEWConfig](/docs/transformers/v5.23.0/ko/model_doc/sew#transformers.SEWConfig) 구성 클래스: [SEWForCTC](/docs/transformers/v5.23.0/ko/model_doc/sew#transformers.SEWForCTC) (SEWConfig 모델)
- [SEWDConfig](/docs/transformers/v5.23.0/ko/model_doc/sew-d#transformers.SEWDConfig) 구성 클래스: [SEWDForCTC](/docs/transformers/v5.23.0/ko/model_doc/sew-d#transformers.SEWDForCTC) (SEWDConfig 모델)
- [UniSpeechConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechConfig) 구성 클래스: [UniSpeechForCTC](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechForCTC) (UniSpeechConfig 모델)
- [UniSpeechSatConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatConfig) 구성 클래스: [UniSpeechSatForCTC](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatForCTC) (UniSpeechSatConfig 모델)
- [Wav2Vec2BertConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertConfig) 구성 클래스: [Wav2Vec2BertForCTC](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertForCTC) (Wav2Vec2BertConfig 모델)
- [Wav2Vec2Config](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Config) 구성 클래스: [Wav2Vec2ForCTC](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2ForCTC) (Wav2Vec2Config 모델)
- [Wav2Vec2ConformerConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerConfig) 구성 클래스: [Wav2Vec2ConformerForCTC](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerForCTC) (Wav2Vec2ConformerConfig 모델)
- [WavLMConfig](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMConfig) 구성 클래스: [WavLMForCTC](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMForCTC) (WavLMConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForCTC

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForCTC.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForCTC.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **data2vec-audio** -- [Data2VecAudioForCTC](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioForCTC) (Data2VecAudioConfig 모델)
- **granite_speech5_ctc** -- [GraniteSpeech5ForCTC](/docs/transformers/v5.23.0/ko/model_doc/granite_speech5#transformers.GraniteSpeech5ForCTC) (GraniteSpeech5CTCConfig 모델)
- **hubert** -- [HubertForCTC](/docs/transformers/v5.23.0/ko/model_doc/hubert#transformers.HubertForCTC) (HubertConfig 모델)
- **lasr_ctc** -- [LasrForCTC](/docs/transformers/v5.23.0/ko/model_doc/lasr#transformers.LasrForCTC) (LasrCTCConfig 모델)
- **parakeet_ctc** -- [ParakeetForCTC](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForCTC) (ParakeetCTCConfig 모델)
- **sew** -- [SEWForCTC](/docs/transformers/v5.23.0/ko/model_doc/sew#transformers.SEWForCTC) (SEWConfig 모델)
- **sew-d** -- [SEWDForCTC](/docs/transformers/v5.23.0/ko/model_doc/sew-d#transformers.SEWDForCTC) (SEWDConfig 모델)
- **unispeech** -- [UniSpeechForCTC](/docs/transformers/v5.23.0/ko/model_doc/unispeech#transformers.UniSpeechForCTC) (UniSpeechConfig 모델)
- **unispeech-sat** -- [UniSpeechSatForCTC](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatForCTC) (UniSpeechSatConfig 모델)
- **wav2vec2** -- [Wav2Vec2ForCTC](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2ForCTC) (Wav2Vec2Config 모델)
- **wav2vec2-bert** -- [Wav2Vec2BertForCTC](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertForCTC) (Wav2Vec2BertConfig 모델)
- **wav2vec2-conformer** -- [Wav2Vec2ConformerForCTC](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerForCTC) (Wav2Vec2ConformerConfig 모델)
- **wavlm** -- [WavLMForCTC](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMForCTC) (WavLMConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForCTC

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForCTC.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForCTC.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForTDT[[transformers.AutoModelForTDT]]

#### transformers.AutoModelForTDT[[transformers.AutoModelForTDT]]

```python
transformers.AutoModelForTDT(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2602)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다 (토큰-기간 트랜스듀서 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForTDT.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 설정 클래스에 따라 선택됩니다:  - [ParakeetTDTConfig](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetTDTConfig) 설정 클래스: [ParakeetForTDT](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForTDT) (ParakeetTDTConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

token-and-duration transducer head를 갖춘 라이브러리의 모델 클래스 중 하나를 구성에서 인스턴스화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTDT

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForTDT.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForTDT.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

token-and-duration transducer head를 갖춘 라이브러리의 모델 클래스 중 하나를 사전 학습된 모델에서 초기화합니다.

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **parakeet_tdt** -- [ParakeetForTDT](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForTDT) (ParakeetTDTConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTDT

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForTDT.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForTDT.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForRNNT[[transformers.AutoModelForRNNT]]

#### transformers.AutoModelForRNNT[[transformers.AutoModelForRNNT]]

```python
transformers.AutoModelForRNNT(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2595)

This is a generic model class that will be instantiated as one of the model classes of the library (with a RNN transducer head) when created with the [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) class method or the [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) class method.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForRNNT.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 설정 클래스에 따라 선택됩니다:
- [Nemotron3_5AsrConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_asr#transformers.Nemotron3_5AsrConfig) 설정 클래스: [Nemotron3_5AsrForRNNT](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_asr#transformers.Nemotron3_5AsrForRNNT) (Nemotron3_5AsrConfig 모델)
- [NemotronAsrStreamingConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingConfig) 설정 클래스: [NemotronAsrStreamingForRNNT](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingForRNNT) (NemotronAsrStreamingConfig 모델)
- [ParakeetRNNTConfig](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetRNNTConfig) 설정 클래스: [ParakeetForRNNT](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForRNNT) (ParakeetRNNTConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForRNNT

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForRNNT.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForRNNT.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **nemotron3_5_asr** -- [Nemotron3_5AsrForRNNT](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_asr#transformers.Nemotron3_5AsrForRNNT) (Nemotron3_5AsrConfig 모델)
- **nemotron_asr_streaming** -- [NemotronAsrStreamingForRNNT](/docs/transformers/v5.23.0/ko/model_doc/nemotron_asr_streaming#transformers.NemotronAsrStreamingForRNNT) (NemotronAsrStreamingConfig 모델)
- **parakeet_rnnt** -- [ParakeetForRNNT](/docs/transformers/v5.23.0/ko/model_doc/parakeet#transformers.ParakeetForRNNT) (ParakeetRNNTConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForRNNT

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForRNNT.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForRNNT.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForSpeechSeq2Seq[[transformers.AutoModelForSpeechSeq2Seq]]

#### transformers.AutoModelForSpeechSeq2Seq[[transformers.AutoModelForSpeechSeq2Seq]]

```python
transformers.AutoModelForSpeechSeq2Seq(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2609)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 시퀀스-투-시퀀스 음성-투-텍스트 모델링 헤드를 갖춘 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 범용 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForSpeechSeq2Seq.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:
- [CanaryConfig](/docs/transformers/v5.23.0/ko/model_doc/canary#transformers.CanaryConfig) 구성 클래스: [CanaryForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/canary#transformers.CanaryForConditionalGeneration) (CanaryConfig 모델)
- [CohereAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_asr#transformers.CohereAsrConfig) 구성 클래스: [CohereAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cohere_asr#transformers.CohereAsrForConditionalGeneration) (CohereAsrConfig 모델)
- [DiaConfig](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaConfig) 구성 클래스: [DiaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaForConditionalGeneration) (DiaConfig 모델)
- [FunAsrNanoConfig](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoConfig) 구성 클래스: [FunAsrNanoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoForConditionalGeneration) (FunAsrNanoConfig 모델)
- [GraniteSpeechConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechConfig) 구성 클래스: [GraniteSpeechForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechForConditionalGeneration) (GraniteSpeechConfig 모델)
- [GraniteSpeechPlusConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusConfig) 구성 클래스: [GraniteSpeechPlusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusForConditionalGeneration) (GraniteSpeechPlusConfig 모델)
- [KyutaiSpeechToTextConfig](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextConfig) 구성 클래스: [KyutaiSpeechToTextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextForConditionalGeneration) (KyutaiSpeechToTextConfig 모델)
- [MoonshineConfig](/docs/transformers/v5.23.0/ko/model_doc/moonshine#transformers.MoonshineConfig) 구성 클래스: [MoonshineForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moonshine#transformers.MoonshineForConditionalGeneration) (MoonshineConfig 모델)
- [MoonshineStreamingConfig](/docs/transformers/v5.23.0/ko/model_doc/moonshine_streaming#transformers.MoonshineStreamingConfig) 구성 클래스: [MoonshineStreamingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moonshine_streaming#transformers.MoonshineStreamingForConditionalGeneration) (MoonshineStreamingConfig 모델)
- [Pop2PianoConfig](/docs/transformers/v5.23.0/ko/model_doc/pop2piano#transformers.Pop2PianoConfig) 구성 클래스: [Pop2PianoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pop2piano#transformers.Pop2PianoForConditionalGeneration) (Pop2PianoConfig 모델)
- [Qwen3ASRConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRConfig) 구성 클래스: [Qwen3ASRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRForConditionalGeneration) (Qwen3ASRConfig 모델)
- [SeamlessM4TConfig](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TConfig) 구성 클래스: [SeamlessM4TForSpeechToText](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TForSpeechToText) (SeamlessM4TConfig 모델)
- [SeamlessM4Tv2Config](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t_v2#transformers.SeamlessM4Tv2Config) 구성 클래스: [SeamlessM4Tv2ForSpeechToText](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t_v2#transformers.SeamlessM4Tv2ForSpeechToText) (SeamlessM4Tv2Config 모델)
- [Speech2TextConfig](/docs/transformers/v5.23.0/ko/model_doc/speech_to_text#transformers.Speech2TextConfig) 구성 클래스: [Speech2TextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/speech_to_text#transformers.Speech2TextForConditionalGeneration) (Speech2TextConfig 모델)
- [SpeechEncoderDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/speech-encoder-decoder#transformers.SpeechEncoderDecoderConfig) 구성 클래스: [SpeechEncoderDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/speech-encoder-decoder#transformers.SpeechEncoderDecoderModel) (SpeechEncoderDecoderConfig 모델)
- [SpeechT5Config](/docs/transformers/v5.23.0/ko/model_doc/speecht5#transformers.SpeechT5Config) 구성 클래스: [SpeechT5ForSpeechToText](/docs/transformers/v5.23.0/ko/model_doc/speecht5#transformers.SpeechT5ForSpeechToText) (SpeechT5Config 모델)
- [VibeVoiceAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrConfig) 구성 클래스: [VibeVoiceAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrForConditionalGeneration) (VibeVoiceAsrConfig 모델)
- [VibeVoiceConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice#transformers.VibeVoiceConfig) 구성 클래스: [VibeVoiceForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice#transformers.VibeVoiceForConditionalGeneration) (VibeVoiceConfig 모델)
- [VoxtralConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralConfig) 구성 클래스: [VoxtralForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralForConditionalGeneration) (VoxtralConfig 모델)
- [VoxtralRealtimeConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeConfig) 구성 클래스: [VoxtralRealtimeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeForConditionalGeneration) (VoxtralRealtimeConfig 모델)
- [WhisperConfig](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperConfig) 구성 클래스: [WhisperForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperForConditionalGeneration) (WhisperConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForSpeechSeq2Seq

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForSpeechSeq2Seq.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForSpeechSeq2Seq.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **canary** -- [CanaryForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/canary#transformers.CanaryForConditionalGeneration) (CanaryConfig 모델)
- **cohere_asr** -- [CohereAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cohere_asr#transformers.CohereAsrForConditionalGeneration) (CohereAsrConfig 모델)
- **dia** -- [DiaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/dia#transformers.DiaForConditionalGeneration) (DiaConfig 모델)
- **fun_asr_nano** -- [FunAsrNanoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoForConditionalGeneration) (FunAsrNanoConfig 모델)
- **granite_speech** -- [GraniteSpeechForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechForConditionalGeneration) (GraniteSpeechConfig 모델)
- **granite_speech_plus** -- [GraniteSpeechPlusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusForConditionalGeneration) (GraniteSpeechPlusConfig 모델)
- **kyutai_speech_to_text** -- [KyutaiSpeechToTextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextForConditionalGeneration) (KyutaiSpeechToTextConfig 모델)
- **moonshine** -- [MoonshineForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moonshine#transformers.MoonshineForConditionalGeneration) (MoonshineConfig 모델)
- **moonshine_streaming** -- [MoonshineStreamingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moonshine_streaming#transformers.MoonshineStreamingForConditionalGeneration) (MoonshineStreamingConfig 모델)
- **pop2piano** -- [Pop2PianoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pop2piano#transformers.Pop2PianoForConditionalGeneration) (Pop2PianoConfig 모델)
- **qwen3_asr** -- [Qwen3ASRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRForConditionalGeneration) (Qwen3ASRConfig 모델)
- **seamless_m4t** -- [SeamlessM4TForSpeechToText](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t#transformers.SeamlessM4TForSpeechToText) (SeamlessM4TConfig 모델)
- **seamless_m4t_v2** -- [SeamlessM4Tv2ForSpeechToText](/docs/transformers/v5.23.0/ko/model_doc/seamless_m4t_v2#transformers.SeamlessM4Tv2ForSpeechToText) (SeamlessM4Tv2Config 모델)
- **speech-encoder-decoder** -- [SpeechEncoderDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/speech-encoder-decoder#transformers.SpeechEncoderDecoderModel) (SpeechEncoderDecoderConfig 모델)
- **speech_to_text** -- [Speech2TextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/speech_to_text#transformers.Speech2TextForConditionalGeneration) (Speech2TextConfig 모델)
- **speecht5** -- [SpeechT5ForSpeechToText](/docs/transformers/v5.23.0/ko/model_doc/speecht5#transformers.SpeechT5ForSpeechToText) (SpeechT5Config 모델)
- **vibevoice** -- [VibeVoiceForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice#transformers.VibeVoiceForConditionalGeneration) (VibeVoiceConfig 모델)
- **vibevoice_asr** -- [VibeVoiceAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrForConditionalGeneration) (VibeVoiceAsrConfig 모델)
- **voxtral** -- [VoxtralForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralForConditionalGeneration) (VoxtralConfig 모델)
- **voxtral_realtime** -- [VoxtralRealtimeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeForConditionalGeneration) (VoxtralRealtimeConfig 모델)
- **whisper** -- [WhisperForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/whisper#transformers.WhisperForConditionalGeneration) (WhisperConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForSpeechSeq2Seq

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForSpeechSeq2Seq.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForSpeechSeq2Seq.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForAudioXVector[[transformers.AutoModelForAudioXVector]]

#### transformers.AutoModelForAudioXVector[[transformers.AutoModelForAudioXVector]]

```python
transformers.AutoModelForAudioXVector(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2627)

이 범용 모델 클래스는 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때 라이브러리의 모델 클래스 중 하나로 인스턴스화됩니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForAudioXVector.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다: - [Data2VecAudioConfig](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioConfig) 구성 클래스: [Data2VecAudioForXVector](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioForXVector) (Data2VecAudioConfig 모델) - [UniSpeechSatConfig](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatConfig) 구성 클래스: [UniSpeechSatForXVector](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatForXVector) (UniSpeechSatConfig 모델) - [Wav2Vec2BertConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertConfig) 구성 클래스: [Wav2Vec2BertForXVector](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertForXVector) (Wav2Vec2BertConfig 모델) - [Wav2Vec2Config](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2Config) 구성 클래스: [Wav2Vec2ForXVector](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2ForXVector) (Wav2Vec2Config 모델) - [Wav2Vec2ConformerConfig](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerConfig) 구성 클래스: [Wav2Vec2ConformerForXVector](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerForXVector) (Wav2Vec2ConformerConfig 모델) - [WavLMConfig](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMConfig) 구성 클래스: [WavLMForXVector](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMForXVector) (WavLMConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

… 라이브러리의 모델 클래스 중 하나(오디오 검색을 x-vector 헤드를 통해 수행하는 모델)를 구성에서 인스턴스화합니다.

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForAudioXVector

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForAudioXVector.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForAudioXVector.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **data2vec-audio** -- [Data2VecAudioForXVector](/docs/transformers/v5.23.0/ko/model_doc/data2vec#transformers.Data2VecAudioForXVector) (Data2VecAudioConfig 모델)
- **unispeech-sat** -- [UniSpeechSatForXVector](/docs/transformers/v5.23.0/ko/model_doc/unispeech-sat#transformers.UniSpeechSatForXVector) (UniSpeechSatConfig 모델)
- **wav2vec2** -- [Wav2Vec2ForXVector](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2#transformers.Wav2Vec2ForXVector) (Wav2Vec2Config 모델)
- **wav2vec2-bert** -- [Wav2Vec2BertForXVector](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-bert#transformers.Wav2Vec2BertForXVector) (Wav2Vec2BertConfig 모델)
- **wav2vec2-conformer** -- [Wav2Vec2ConformerForXVector](/docs/transformers/v5.23.0/ko/model_doc/wav2vec2-conformer#transformers.Wav2Vec2ConformerForXVector) (Wav2Vec2ConformerConfig 모델)
- **wavlm** -- [WavLMForXVector](/docs/transformers/v5.23.0/ko/model_doc/wavlm#transformers.WavLMForXVector) (WavLMConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForAudioXVector

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForAudioXVector.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForAudioXVector.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForTextToSpectrogram[[transformers.AutoModelForTextToSpectrogram]]

#### transformers.AutoModelForTextToSpectrogram[[transformers.AutoModelForTextToSpectrogram]]

```python
transformers.AutoModelForTextToSpectrogram(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2631)

### AutoModelForTextToWaveform[[transformers.AutoModelForTextToWaveform]]

#### transformers.AutoModelForTextToWaveform[[transformers.AutoModelForTextToWaveform]]

```python
transformers.AutoModelForTextToWaveform(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2635)

### AutoModelForAudioTokenization[[transformers.AutoModelForAudioTokenization]]

#### transformers.AutoModelForAudioTokenization[[transformers.AutoModelForAudioTokenization]]

```python
transformers.AutoModelForAudioTokenization(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2653)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 범용 모델 클래스입니다 (codebooks 헤드를 통한 오디오 토큰화 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForAudioTokenization.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다:
- [DacConfig](/docs/transformers/v5.23.0/ko/model_doc/dac#transformers.DacConfig) 구성 클래스: [DacModel](/docs/transformers/v5.23.0/ko/model_doc/dac#transformers.DacModel) (DacConfig 모델)
- [HiggsAudioV2TokenizerConfig](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2_tokenizer#transformers.HiggsAudioV2TokenizerConfig) 구성 클래스: [HiggsAudioV2TokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2_tokenizer#transformers.HiggsAudioV2TokenizerModel) (HiggsAudioV2TokenizerConfig 모델)
- [NeuCodecConfig](/docs/transformers/v5.23.0/ko/model_doc/neucodec#transformers.NeuCodecConfig) 구성 클래스: [NeuCodecModel](/docs/transformers/v5.23.0/ko/model_doc/neucodec#transformers.NeuCodecModel) (NeuCodecConfig 모델)
- [PocketTTSAudioTokenizerConfig](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerConfig) 구성 클래스: [PocketTTSAudioTokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerModel) (PocketTTSAudioTokenizerConfig 모델)
- [VibeVoiceAcousticTokenizerConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerConfig) 구성 클래스: [VibeVoiceAcousticTokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerModel) (VibeVoiceAcousticTokenizerConfig 모델)
- [Xcodec2Config](/docs/transformers/v5.23.0/ko/model_doc/xcodec2#transformers.Xcodec2Config) 구성 클래스: [Xcodec2Model](/docs/transformers/v5.23.0/ko/model_doc/xcodec2#transformers.Xcodec2Model) (Xcodec2Config 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForAudioTokenization

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForAudioTokenization.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForAudioTokenization.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **dac** -- [DacModel](/docs/transformers/v5.23.0/ko/model_doc/dac#transformers.DacModel) (DacConfig 모델)
- **higgs_audio_v2_tokenizer** -- [HiggsAudioV2TokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/higgs_audio_v2_tokenizer#transformers.HiggsAudioV2TokenizerModel) (HiggsAudioV2TokenizerConfig 모델)
- **neucodec** -- [NeuCodecModel](/docs/transformers/v5.23.0/ko/model_doc/neucodec#transformers.NeuCodecModel) (NeuCodecConfig 모델)
- **pocket_tts_audio_tokenizer** -- [PocketTTSAudioTokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/pocket_tts_audio_tokenizer#transformers.PocketTTSAudioTokenizerModel) (PocketTTSAudioTokenizerConfig 모델)
- **vibevoice_acoustic_tokenizer** -- [VibeVoiceAcousticTokenizerModel](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_acoustic_tokenizer#transformers.VibeVoiceAcousticTokenizerModel) (VibeVoiceAcousticTokenizerConfig 모델)
- **xcodec2** -- [Xcodec2Model](/docs/transformers/v5.23.0/ko/model_doc/xcodec2#transformers.Xcodec2Model) (Xcodec2Config 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForAudioTokenization

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForAudioTokenization.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForAudioTokenization.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

## 멀티모달

다음과 같은 멀티모달 작업을 위해 다음 auto 클래스가 사용할 수 있습니다.

### AutoModelForMultimodalLM[[transformers.AutoModelForMultimodalLM]]

#### transformers.AutoModelForMultimodalLM[[transformers.AutoModelForMultimodalLM]]

```python
transformers.AutoModelForMultimodalLM(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2574)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 범용 모델 클래스입니다(멀티모달 생성 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForMultimodalLM.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : The model class to instantiate is selected based on the configuration class:  - [AriaConfig](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaConfig) configuration class: [AriaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaForConditionalGeneration) (AriaConfig model) - [AyaVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/aya_vision#transformers.AyaVisionConfig) configuration class: [AyaVisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/aya_vision#transformers.AyaVisionForConditionalGeneration) (AyaVisionConfig model) - [Blip2Config](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2Config) configuration class: [Blip2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2ForConditionalGeneration) (Blip2Config model) - [BlipConfig](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipConfig) configuration class: [BlipForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipForConditionalGeneration) (BlipConfig model) - [ChameleonConfig](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonConfig) configuration class: [ChameleonForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonForConditionalGeneration) (ChameleonConfig model) - [Cohere2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere2_vision#transformers.Cohere2VisionConfig) configuration class: [Cohere2VisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cohere2_vision#transformers.Cohere2VisionForConditionalGeneration) (Cohere2VisionConfig model) - [CohereCompassConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassConfig) configuration class: [CohereCompassForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassForConditionalGeneration) (CohereCompassConfig model) - [Cosmos3EdgeConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeConfig) configuration class: [Cosmos3EdgeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeForConditionalGeneration) (Cosmos3EdgeConfig model) - [Cosmos3OmniConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_omni#transformers.Cosmos3OmniConfig) configuration class: [Cosmos3OmniForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_omni#transformers.Cosmos3OmniForConditionalGeneration) (Cosmos3OmniConfig model) - [DeepseekOcr2Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2Config) configuration class: [DeepseekOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2ForConditionalGeneration) (DeepseekOcr2Config model) - [DeepseekVLConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl#transformers.DeepseekVLConfig) configuration class: [DeepseekVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl#transformers.DeepseekVLForConditionalGeneration) (DeepseekVLConfig model) - [DeepseekVLHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl_hybrid#transformers.DeepseekVLHybridConfig) configuration class: [DeepseekVLHybridForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl_hybrid#transformers.DeepseekVLHybridForConditionalGeneration) (DeepseekVLHybridConfig model) - [DiffusionGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/diffusion_gemma#transformers.DiffusionGemmaConfig) configuration class: [DiffusionGemmaForBlockDiffusion](/docs/transformers/v5.23.0/ko/model_doc/diffusion_gemma#transformers.DiffusionGemmaForBlockDiffusion) (DiffusionGemmaConfig model) - [Dots3NoteConfig](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteConfig) configuration class: [Dots3NoteForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteForConditionalGeneration) (Dots3NoteConfig model) - [Emu3Config](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3Config) configuration class: [Emu3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3ForConditionalGeneration) (Emu3Config model) - [Ernie4_5_VLMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeConfig) configuration class: [Ernie4_5_VLMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeForConditionalGeneration) (Ernie4_5_VLMoeConfig model) - [EvollaConfig](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaConfig) configuration class: [EvollaForProteinText2Text](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaForProteinText2Text) (EvollaConfig model) - [Exaone4_5_Config](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_Config) configuration class: [Exaone4_5_ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_ForConditionalGeneration) (Exaone4_5_Config model) - [FastVlmConfig](/docs/transformers/v5.23.0/ko/model_doc/fast_vlm#transformers.FastVlmConfig) configuration class: [FastVlmForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fast_vlm#transformers.FastVlmForConditionalGeneration) (FastVlmConfig model) - [Florence2Config](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2Config) configuration class: [Florence2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2ForConditionalGeneration) (Florence2Config model) - [FunAsrNanoConfig](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoConfig) configuration class: [FunAsrNanoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoForConditionalGeneration) (FunAsrNanoConfig model) - [FuyuConfig](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuConfig) configuration class: [FuyuForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuForCausalLM) (FuyuConfig model) - [Gemma3Config](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Config) configuration class: [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (Gemma3Config model) - [Gemma3nConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nConfig) configuration class: [Gemma3nForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nForConditionalGeneration) (Gemma3nConfig model) - [Gemma4Config](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4Config) configuration class: [Gemma4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4ForConditionalGeneration) (Gemma4Config model) - [Gemma4UnifiedConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedConfig) configuration class: [Gemma4UnifiedForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedForConditionalGeneration) (Gemma4UnifiedConfig model) - [GitConfig](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitConfig) configuration class: [GitForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitForCausalLM) (GitConfig model) - [Glm46VConfig](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VConfig) configuration class: [Glm46VForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VForConditionalGeneration) (Glm46VConfig model) - [Glm4vConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vConfig) configuration class: [Glm4vForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vForConditionalGeneration) (Glm4vConfig model) - [Glm4vMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeConfig) configuration class: [Glm4vMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeForConditionalGeneration) (Glm4vMoeConfig model) - [Glm5NextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextConfig) configuration class: [Glm5NextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextForConditionalGeneration) (Glm5NextConfig model) - [GlmAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrConfig) configuration class: [GlmAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrForConditionalGeneration) (GlmAsrConfig model) - [GlmOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrConfig) configuration class: [GlmOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrForConditionalGeneration) (GlmOcrConfig model) - [GlmgaConfig](/docs/transformers/v5.23.0/ko/model_doc/glmga#transformers.GlmgaConfig) configuration class: [Glm46VForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VForConditionalGeneration) (GlmgaConfig model) - [GotOcr2Config](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2Config) configuration class: [GotOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2ForConditionalGeneration) (GotOcr2Config model) - [Granite4VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionConfig) configuration class: [Granite4VisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionForConditionalGeneration) (Granite4VisionConfig model) - [GraniteDoclingConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingConfig) configuration class: [GraniteDoclingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingForConditionalGeneration) (GraniteDoclingConfig model) - [GraniteSpeechConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechConfig) configuration class: [GraniteSpeechForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechForConditionalGeneration) (GraniteSpeechConfig model) - [GraniteSpeechPlusConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusConfig) configuration class: [GraniteSpeechPlusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusForConditionalGeneration) (GraniteSpeechPlusConfig model) - [HunYuanVLConfig](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLConfig) configuration class: [HunYuanVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLForConditionalGeneration) (HunYuanVLConfig model) - [HyperCLOVAXVisionV2Config](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax_vision_v2#transformers.HyperCLOVAXVisionV2Config) configuration class: [HyperCLOVAXVisionV2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax_vision_v2#transformers.HyperCLOVAXVisionV2ForConditionalGeneration) (HyperCLOVAXVisionV2Config model) - [Idefics2Config](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2Config) configuration class: [Idefics2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2ForConditionalGeneration) (Idefics2Config model) - [Idefics3Config](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3Config) configuration class: [Idefics3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3ForConditionalGeneration) (Idefics3Config model) - [IdeficsConfig](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsConfig) configuration class: [IdeficsForVisionText2Text](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsForVisionText2Text) (IdeficsConfig model) - [Inkling2Config](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2Config) configuration class: [Inkling2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2ForConditionalGeneration) (Inkling2Config model) - [InklingConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingConfig) configuration class: [InklingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingForConditionalGeneration) (InklingConfig model) - [InstructBlipConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipConfig) configuration class: [InstructBlipForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipForConditionalGeneration) (InstructBlipConfig model) - [InstructBlipVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoConfig) configuration class: [InstructBlipVideoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoForConditionalGeneration) (InstructBlipVideoConfig model) - [InternVLConfig](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLConfig) configuration class: [InternVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLForConditionalGeneration) (InternVLConfig model) - [JanusConfig](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusConfig) configuration class: [JanusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusForConditionalGeneration) (JanusConfig model) - [Kimi_K25Config](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25Config) configuration class: [Kimi_K25ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25ForConditionalGeneration) (Kimi_K25Config model) - [Kosmos2Config](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2Config) configuration class: [Kosmos2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2ForConditionalGeneration) (Kosmos2Config model) - [Kosmos2_5Config](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5Config) configuration class: [Kosmos2_5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5ForConditionalGeneration) (Kosmos2_5Config model) - [KyutaiSpeechToTextConfig](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextConfig) configuration class: [KyutaiSpeechToTextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextForConditionalGeneration) (KyutaiSpeechToTextConfig model) - [Lfm2VlConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm2_vl#transformers.Lfm2VlConfig) configuration class: [Lfm2VlForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lfm2_vl#transformers.Lfm2VlForConditionalGeneration) (Lfm2VlConfig model) - [Lfm3VlConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlConfig) configuration class: [Lfm3VlForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlForConditionalGeneration) (Lfm3VlConfig model) - [LightOnOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/lighton_ocr#transformers.LightOnOcrConfig) configuration class: [LightOnOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lighton_ocr#transformers.LightOnOcrForConditionalGeneration) (LightOnOcrConfig model) - [Llama4Config](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4Config) configuration class: [Llama4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4ForConditionalGeneration) (Llama4Config model) - [LlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaConfig) configuration class: [LlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaForConditionalGeneration) (LlavaConfig model) - [LlavaNextConfig](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextConfig) configuration class: [LlavaNextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextForConditionalGeneration) (LlavaNextConfig model) - [LlavaNextVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoConfig) configuration class: [LlavaNextVideoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoForConditionalGeneration) (LlavaNextVideoConfig model) - [LlavaOnevisionConfig](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionConfig) configuration class: [LlavaOnevisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionForConditionalGeneration) (LlavaOnevisionConfig model) - [MiniCPMOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrConfig) configuration class: [MiniCPMOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrForConditionalGeneration) (MiniCPMOcrConfig model) - [MiniCPMV4_6Config](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6Config) configuration class: [MiniCPMV4_6ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6ForConditionalGeneration) (MiniCPMV4_6Config model) - [MiniCPMV4_7Config](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7Config) configuration class: [MiniCPMV4_7ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7ForConditionalGeneration) (MiniCPMV4_7Config model) - [MiniMaxM3VLConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLConfig) configuration class: [MiniMaxM3SparseForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3SparseForConditionalGeneration) (MiniMaxM3VLConfig model) - [Mistral3Config](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3Config) configuration class: [Mistral3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3ForConditionalGeneration) (Mistral3Config model) - [Mistral4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4Config) configuration class: [Mistral4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4ForCausalLM) (Mistral4Config model) - [MistralLarge4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4Config) configuration class: [MistralLarge4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4ForConditionalGeneration) (MistralLarge4Config model) - [MllamaConfig](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaConfig) configuration class: [MllamaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaForConditionalGeneration) (MllamaConfig model) - [Molmo2Config](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2Config) configuration class: [Molmo2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2ForConditionalGeneration) (Molmo2Config model) - [MossTranscribeDiarizeConfig](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeConfig) configuration class: [MossTranscribeDiarizeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeForConditionalGeneration) (MossTranscribeDiarizeConfig model) - [MuseGlimmerConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerConfig) configuration class: [MuseGlimmerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerForConditionalGeneration) (MuseGlimmerConfig model) - [MuseSparkConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkConfig) configuration class: [MuseSparkForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkForConditionalGeneration) (MuseSparkConfig model) - [Nemotron3_5TranscribeConfig](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeConfig) configuration class: [Nemotron3_5TranscribeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeForConditionalGeneration) (Nemotron3_5TranscribeConfig model) - [NemotronH_Omni_Reasoning_V3_Config](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h_omni#transformers.NemotronH_Omni_Reasoning_V3_Config) configuration class: [NemotronH_Omni_Reasoning_V3](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h_omni#transformers.NemotronH_Omni_Reasoning_V3) (NemotronH_Omni_Reasoning_V3_Config model) - [Ovis2Config](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2Config) configuration class: [Ovis2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2ForConditionalGeneration) (Ovis2Config model) - [PI0Config](/docs/transformers/v5.23.0/ko/model_doc/pi0#transformers.PI0Config) configuration class: [PI0ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pi0#transformers.PI0ForConditionalGeneration) (PI0Config model) - [PPChart2TableConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_chart2table#transformers.PPChart2TableConfig) configuration class: [GotOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2ForConditionalGeneration) (PPChart2TableConfig model) - [PPFormulaNetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_formulanet#transformers.PPFormulaNetConfig) configuration class: [PPFormulaNetForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pp_formulanet#transformers.PPFormulaNetForConditionalGeneration) (PPFormulaNetConfig model) - [PaddleOCRVLConfig](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVLConfig) configuration class: [PaddleOCRVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVLForConditionalGeneration) (PaddleOCRVLConfig model) - [PaliGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaConfig) configuration class: [PaliGemmaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaForConditionalGeneration) (PaliGemmaConfig model) - [PerceptionLMConfig](/docs/transformers/v5.23.0/ko/model_doc/perception_lm#transformers.PerceptionLMConfig) configuration class: [PerceptionLMForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/perception_lm#transformers.PerceptionLMForConditionalGeneration) (PerceptionLMConfig model) - [Phi4MultimodalConfig](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalConfig) configuration class: [Phi4MultimodalForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalForCausalLM) (Phi4MultimodalConfig model) - [Pix2StructConfig](/docs/transformers/v5.23.0/ko/model_doc/pix2struct#transformers.Pix2StructConfig) configuration class: [Pix2StructForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pix2struct#transformers.Pix2StructForConditionalGeneration) (Pix2StructConfig model) - [QianfanOCRConfig](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRConfig) configuration class: [QianfanOCRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRForConditionalGeneration) (QianfanOCRConfig model) - [Qwen2AudioConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioConfig) configuration class: [Qwen2AudioForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioForConditionalGeneration) (Qwen2AudioConfig model) - [Qwen2VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLConfig) configuration class: [Qwen2VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLForConditionalGeneration) (Qwen2VLConfig model) - [Qwen2_5OmniConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniConfig) configuration class: [Qwen2_5OmniForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniForConditionalGeneration) (Qwen2_5OmniConfig model) - [Qwen2_5OmniThinkerConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniThinkerConfig) configuration class: [Qwen2_5OmniThinkerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniThinkerForConditionalGeneration) (Qwen2_5OmniThinkerConfig model) - [Qwen2_5_VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLConfig) configuration class: [Qwen2_5_VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLForConditionalGeneration) (Qwen2_5_VLConfig model) - [Qwen3ASRConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRConfig) configuration class: [Qwen3ASRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRForConditionalGeneration) (Qwen3ASRConfig model) - [Qwen3OmniMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeConfig) configuration class: [Qwen3OmniMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeForConditionalGeneration) (Qwen3OmniMoeConfig model) - [Qwen3OmniMoeThinkerConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeThinkerConfig) configuration class: [Qwen3OmniMoeThinkerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeThinkerForConditionalGeneration) (Qwen3OmniMoeThinkerConfig model) - [Qwen3VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLConfig) configuration class: [Qwen3VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLForConditionalGeneration) (Qwen3VLConfig model) - [Qwen3VLMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeConfig) configuration class: [Qwen3VLMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeForConditionalGeneration) (Qwen3VLMoeConfig model) - [Qwen3_5Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Config) configuration class: [Qwen3_5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForConditionalGeneration) (Qwen3_5Config model) - [Qwen3_5MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeConfig) configuration class: [Qwen3_5MoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeForConditionalGeneration) (Qwen3_5MoeConfig model) - [Qwen4ExpConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpConfig) configuration class: [Qwen4ExpForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpForConditionalGeneration) (Qwen4ExpConfig model) - [ShieldGemma2Config](/docs/transformers/v5.23.0/ko/model_doc/shieldgemma2#transformers.ShieldGemma2Config) configuration class: [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (ShieldGemma2Config model) - [SmolVLMConfig](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMConfig) configuration class: [SmolVLMForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMForConditionalGeneration) (SmolVLMConfig model) - [Step3p7Config](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7Config) configuration class: [Step3p7ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7ForConditionalGeneration) (Step3p7Config model) - [T5Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2Config) configuration class: [T5Gemma2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForConditionalGeneration) (T5Gemma2Config model) - [UdopConfig](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopConfig) configuration class: [UdopForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopForConditionalGeneration) (UdopConfig model) - [UnlimitedOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrConfig) configuration class: [UnlimitedOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrForConditionalGeneration) (UnlimitedOcrConfig model) - [VibeVoiceAsrConfig](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrConfig) configuration class: [VibeVoiceAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrForConditionalGeneration) (VibeVoiceAsrConfig model) - [VideoLlama3Config](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3Config) configuration class: [VideoLlama3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3ForConditionalGeneration) (VideoLlama3Config model) - [VideoLlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaConfig) configuration class: [VideoLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaForConditionalGeneration) (VideoLlavaConfig model) - [VipLlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaConfig) configuration class: [VipLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaForConditionalGeneration) (VipLlavaConfig model) - [VisionEncoderDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/vision-encoder-decoder#transformers.VisionEncoderDecoderConfig) configuration class: [VisionEncoderDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/vision-encoder-decoder#transformers.VisionEncoderDecoderModel) (VisionEncoderDecoderConfig model) - [VoxtralConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralConfig) configuration class: [VoxtralForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralForConditionalGeneration) (VoxtralConfig model) - [VoxtralRealtimeConfig](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeConfig) configuration class: [VoxtralRealtimeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeForConditionalGeneration) (VoxtralRealtimeConfig model)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForMultimodalLM

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForMultimodalLM.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForMultimodalLM.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **aria** -- [AriaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaForConditionalGeneration) (AriaConfig 모델)
- **aya_vision** -- [AyaVisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/aya_vision#transformers.AyaVisionForConditionalGeneration) (AyaVisionConfig 모델)
- **blip** -- [BlipForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipForConditionalGeneration) (BlipConfig 모델)
- **blip-2** -- [Blip2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2ForConditionalGeneration) (Blip2Config 모델)
- **chameleon** -- [ChameleonForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonForConditionalGeneration) (ChameleonConfig 모델)
- **cohere2_vision** -- [Cohere2VisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cohere2_vision#transformers.Cohere2VisionForConditionalGeneration) (Cohere2VisionConfig 모델)
- **cohere_compass** -- [CohereCompassForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassForConditionalGeneration) (CohereCompassConfig 모델)
- **cosmos3_edge** -- [Cosmos3EdgeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeForConditionalGeneration) (Cosmos3EdgeConfig 모델)
- **cosmos3_omni** -- [Cosmos3OmniForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_omni#transformers.Cosmos3OmniForConditionalGeneration) (Cosmos3OmniConfig 모델)
- **deepseek_ocr2** -- [DeepseekOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2ForConditionalGeneration) (DeepseekOcr2Config 모델)
- **deepseek_vl** -- [DeepseekVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl#transformers.DeepseekVLForConditionalGeneration) (DeepseekVLConfig 모델)
- **deepseek_vl_hybrid** -- [DeepseekVLHybridForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl_hybrid#transformers.DeepseekVLHybridForConditionalGeneration) (DeepseekVLHybridConfig 모델)
- **diffusion_gemma** -- [DiffusionGemmaForBlockDiffusion](/docs/transformers/v5.23.0/ko/model_doc/diffusion_gemma#transformers.DiffusionGemmaForBlockDiffusion) (DiffusionGemmaConfig 모델)
- **dots3_note** -- [Dots3NoteForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteForConditionalGeneration) (Dots3NoteConfig 모델)
- **emu3** -- [Emu3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3ForConditionalGeneration) (Emu3Config 모델)
- **ernie4_5_vl_moe** -- [Ernie4_5_VLMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeForConditionalGeneration) (Ernie4_5_VLMoeConfig 모델)
- **evolla** -- [EvollaForProteinText2Text](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaForProteinText2Text) (EvollaConfig 모델)
- **exaone4_5** -- [Exaone4_5_ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_ForConditionalGeneration) (Exaone4_5_Config 모델)
- **fast_vlm** -- [FastVlmForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fast_vlm#transformers.FastVlmForConditionalGeneration) (FastVlmConfig 모델)
- **florence2** -- [Florence2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2ForConditionalGeneration) (Florence2Config 모델)
- **fun_asr_nano** -- [FunAsrNanoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fun_asr_nano#transformers.FunAsrNanoForConditionalGeneration) (FunAsrNanoConfig 모델)
- **fuyu** -- [FuyuForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuForCausalLM) (FuyuConfig 모델)
- **gemma3** -- [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (Gemma3Config 모델)
- **gemma3n** -- [Gemma3nForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nForConditionalGeneration) (Gemma3nConfig 모델)
- **gemma4** -- [Gemma4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4ForConditionalGeneration) (Gemma4Config 모델)
- **gemma4_unified** -- [Gemma4UnifiedForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedForConditionalGeneration) (Gemma4UnifiedConfig 모델)
- **git** -- [GitForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitForCausalLM) (GitConfig 모델)
- **glm46v** -- [Glm46VForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VForConditionalGeneration) (Glm46VConfig 모델)
- **glm4v** -- [Glm4vForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vForConditionalGeneration) (Glm4vConfig 모델)
- **glm4v_moe** -- [Glm4vMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeForConditionalGeneration) (Glm4vMoeConfig 모델)
- **glm5_next** -- [Glm5NextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextForConditionalGeneration) (Glm5NextConfig 모델)
- **glm_ocr** -- [GlmOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrForConditionalGeneration) (GlmOcrConfig 모델)
- **glmasr** -- [GlmAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glmasr#transformers.GlmAsrForConditionalGeneration) (GlmAsrConfig 모델)
- **glmga** -- [Glm46VForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VForConditionalGeneration) (GlmgaConfig 모델)
- **got_ocr2** -- [GotOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2ForConditionalGeneration) (GotOcr2Config 모델)
- **granite4_vision** -- [Granite4VisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionForConditionalGeneration) (Granite4VisionConfig 모델)
- **granite_docling** -- [GraniteDoclingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingForConditionalGeneration) (GraniteDoclingConfig 모델)
- **granite_speech** -- [GraniteSpeechForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech#transformers.GraniteSpeechForConditionalGeneration) (GraniteSpeechConfig 모델)
- **granite_speech_plus** -- [GraniteSpeechPlusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_speech_plus#transformers.GraniteSpeechPlusForConditionalGeneration) (GraniteSpeechPlusConfig 모델)
- **hunyuan_vl** -- [HunYuanVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLForConditionalGeneration) (HunYuanVLConfig 모델)
- **hyperclovax_vision_v2** -- [HyperCLOVAXVisionV2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax_vision_v2#transformers.HyperCLOVAXVisionV2ForConditionalGeneration) (HyperCLOVAXVisionV2Config 모델)
- **idefics** -- [IdeficsForVisionText2Text](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsForVisionText2Text) (IdeficsConfig 모델)
- **idefics2** -- [Idefics2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2ForConditionalGeneration) (Idefics2Config 모델)
- **idefics3** -- [Idefics3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3ForConditionalGeneration) (Idefics3Config 모델)
- **inkling2** -- [Inkling2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2ForConditionalGeneration) (Inkling2Config 모델)
- **inkling_mm_model** -- [InklingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingForConditionalGeneration) (InklingConfig 모델)
- **instructblip** -- [InstructBlipForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipForConditionalGeneration) (InstructBlipConfig 모델)
- **instructblipvideo** -- [InstructBlipVideoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoForConditionalGeneration) (InstructBlipVideoConfig 모델)
- **internvl** -- [InternVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLForConditionalGeneration) (InternVLConfig 모델)
- **janus** -- [JanusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusForConditionalGeneration) (JanusConfig 모델)
- **kimi_k25** -- [Kimi_K25ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25ForConditionalGeneration) (Kimi_K25Config 모델)
- **kosmos-2** -- [Kosmos2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2ForConditionalGeneration) (Kosmos2Config 모델)
- **kosmos-2.5** -- [Kosmos2_5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5ForConditionalGeneration) (Kosmos2_5Config 모델)
- **kyutai_speech_to_text** -- [KyutaiSpeechToTextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kyutai_speech_to_text#transformers.KyutaiSpeechToTextForConditionalGeneration) (KyutaiSpeechToTextConfig 모델)
- **lfm2_vl** -- [Lfm2VlForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lfm2_vl#transformers.Lfm2VlForConditionalGeneration) (Lfm2VlConfig 모델)
- **lfm3_vl** -- [Lfm3VlForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlForConditionalGeneration) (Lfm3VlConfig 모델)
- **lighton_ocr** -- [LightOnOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lighton_ocr#transformers.LightOnOcrForConditionalGeneration) (LightOnOcrConfig 모델)
- **llama4** -- [Llama4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4ForConditionalGeneration) (Llama4Config 모델)
- **llava** -- [LlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaForConditionalGeneration) (LlavaConfig 모델)
- **llava_next** -- [LlavaNextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextForConditionalGeneration) (LlavaNextConfig 모델)
- **llava_next_video** -- [LlavaNextVideoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoForConditionalGeneration) (LlavaNextVideoConfig 모델)
- **llava_onevision** -- [LlavaOnevisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionForConditionalGeneration) (LlavaOnevisionConfig 모델)
- **minicpm_ocr** -- [MiniCPMOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrForConditionalGeneration) (MiniCPMOcrConfig 모델)
- **minicpmv4_6** -- [MiniCPMV4_6ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6ForConditionalGeneration) (MiniCPMV4_6Config 모델)
- **minicpmv4_7** -- [MiniCPMV4_7ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7ForConditionalGeneration) (MiniCPMV4_7Config 모델)
- **minimax_m3_vl** -- [MiniMaxM3SparseForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3SparseForConditionalGeneration) (MiniMaxM3VLConfig 모델)
- **mistral3** -- [Mistral3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3ForConditionalGeneration) (Mistral3Config 모델)
- **mistral4** -- [Mistral4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4ForCausalLM) (Mistral4Config 모델)
- **mistral_large4** -- [MistralLarge4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4ForConditionalGeneration) (MistralLarge4Config 모델)
- **mllama** -- [MllamaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaForConditionalGeneration) (MllamaConfig 모델)
- **molmo2** -- [Molmo2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2ForConditionalGeneration) (Molmo2Config 모델)
- **moss_transcribe_diarize** -- [MossTranscribeDiarizeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/moss_transcribe_diarize#transformers.MossTranscribeDiarizeForConditionalGeneration) (MossTranscribeDiarizeConfig 모델)
- **muse_glimmer** -- [MuseGlimmerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerForConditionalGeneration) (MuseGlimmerConfig 모델)
- **muse_spark** -- [MuseSparkForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkForConditionalGeneration) (MuseSparkConfig 모델)
- **nemo_speechlm** -- [Nemotron3_5TranscribeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/nemotron3_5_transcribe#transformers.Nemotron3_5TranscribeForConditionalGeneration) (Nemotron3_5TranscribeConfig 모델)
- **nemotron_h_omni** -- [NemotronH_Omni_Reasoning_V3](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h_omni#transformers.NemotronH_Omni_Reasoning_V3) (NemotronH_Omni_Reasoning_V3_Config 모델)
- **ovis2** -- [Ovis2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2ForConditionalGeneration) (Ovis2Config 모델)
- **paddleocr_vl** -- [PaddleOCRVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVLForConditionalGeneration) (PaddleOCRVLConfig 모델)
- **paligemma** -- [PaliGemmaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaForConditionalGeneration) (PaliGemmaConfig 모델)
- **perception_lm** -- [PerceptionLMForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/perception_lm#transformers.PerceptionLMForConditionalGeneration) (PerceptionLMConfig 모델)
- **phi4_multimodal** -- [Phi4MultimodalForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/phi4_multimodal#transformers.Phi4MultimodalForCausalLM) (Phi4MultimodalConfig 모델)
- **pi0** -- [PI0ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pi0#transformers.PI0ForConditionalGeneration) (PI0Config 모델)
- **pix2struct** -- [Pix2StructForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pix2struct#transformers.Pix2StructForConditionalGeneration) (Pix2StructConfig 모델)
- **pp_chart2table** -- [GotOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2ForConditionalGeneration) (PPChart2TableConfig 모델)
- **pp_formulanet** -- [PPFormulaNetForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pp_formulanet#transformers.PPFormulaNetForConditionalGeneration) (PPFormulaNetConfig 모델)
- **qianfan_ocr** -- [QianfanOCRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRForConditionalGeneration) (QianfanOCRConfig 모델)
- **qwen2_5_omni** -- [Qwen2_5OmniForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniForConditionalGeneration) (Qwen2_5OmniConfig 모델)
- **qwen2_5_omni_thinker** -- [Qwen2_5OmniThinkerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniThinkerForConditionalGeneration) (Qwen2_5OmniThinkerConfig 모델)
- **qwen2_5_vl** -- [Qwen2_5_VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLForConditionalGeneration) (Qwen2_5_VLConfig 모델)
- **qwen2_audio** -- [Qwen2AudioForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_audio#transformers.Qwen2AudioForConditionalGeneration) (Qwen2AudioConfig 모델)
- **qwen2_vl** -- [Qwen2VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLForConditionalGeneration) (Qwen2VLConfig 모델)
- **qwen3_5** -- [Qwen3_5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForConditionalGeneration) (Qwen3_5Config 모델)
- **qwen3_5_moe** -- [Qwen3_5MoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeForConditionalGeneration) (Qwen3_5MoeConfig 모델)
- **qwen3_asr** -- [Qwen3ASRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_asr#transformers.Qwen3ASRForConditionalGeneration) (Qwen3ASRConfig 모델)
- **qwen3_omni_moe** -- [Qwen3OmniMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeForConditionalGeneration) (Qwen3OmniMoeConfig 모델)
- **qwen3_omni_moe_thinker** -- [Qwen3OmniMoeThinkerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeThinkerForConditionalGeneration) (Qwen3OmniMoeThinkerConfig 모델)
- **qwen3_vl** -- [Qwen3VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLForConditionalGeneration) (Qwen3VLConfig 모델)
- **qwen3_vl_moe** -- [Qwen3VLMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeForConditionalGeneration) (Qwen3VLMoeConfig 모델)
- **qwen4_exp** -- [Qwen4ExpForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpForConditionalGeneration) (Qwen4ExpConfig 모델)
- **shieldgemma2** -- [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (ShieldGemma2Config 모델)
- **smolvlm** -- [SmolVLMForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMForConditionalGeneration) (SmolVLMConfig 모델)
- **step3p7** -- [Step3p7ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7ForConditionalGeneration) (Step3p7Config 모델)
- **t5gemma2** -- [T5Gemma2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForConditionalGeneration) (T5Gemma2Config 모델)
- **udop** -- [UdopForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopForConditionalGeneration) (UdopConfig 모델)
- **unlimited_ocr** -- [UnlimitedOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrForConditionalGeneration) (UnlimitedOcrConfig 모델)
- **vibevoice_asr** -- [VibeVoiceAsrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vibevoice_asr#transformers.VibeVoiceAsrForConditionalGeneration) (VibeVoiceAsrConfig 모델)
- **video_llama_3** -- [VideoLlama3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3ForConditionalGeneration) (VideoLlama3Config 모델)
- **video_llava** -- [VideoLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaForConditionalGeneration) (VideoLlavaConfig 모델)
- **vipllava** -- [VipLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaForConditionalGeneration) (VipLlavaConfig 모델)
- **vision-encoder-decoder** -- [VisionEncoderDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/vision-encoder-decoder#transformers.VisionEncoderDecoderModel) (VisionEncoderDecoderConfig 모델)
- **voxtral** -- [VoxtralForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral#transformers.VoxtralForConditionalGeneration) (VoxtralConfig 모델)
- **voxtral_realtime** -- [VoxtralRealtimeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/voxtral_realtime#transformers.VoxtralRealtimeForConditionalGeneration) (VoxtralRealtimeConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForMultimodalLM

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForMultimodalLM.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForMultimodalLM.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForTableQuestionAnswering[[transformers.AutoModelForTableQuestionAnswering]]

#### transformers.AutoModelForTableQuestionAnswering[[transformers.AutoModelForTableQuestionAnswering]]

```python
transformers.AutoModelForTableQuestionAnswering(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2398)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나(테이블 질의응답 헤드를 포함)로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForTableQuestionAnswering.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다: - [TapasConfig](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasConfig) 구성 클래스: [TapasForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasForQuestionAnswering) (TapasConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTableQuestionAnswering

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google/tapas-base-finetuned-wtq")
>>> model = AutoModelForTableQuestionAnswering.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForTableQuestionAnswering.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **tapas** -- [TapasForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/tapas#transformers.TapasForQuestionAnswering) (TapasConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTableQuestionAnswering

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForTableQuestionAnswering.from_pretrained("google/tapas-base-finetuned-wtq")

>>> # Update configuration during loading
>>> model = AutoModelForTableQuestionAnswering.from_pretrained("google/tapas-base-finetuned-wtq", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForDocumentQuestionAnswering[[transformers.AutoModelForDocumentQuestionAnswering]]

#### transformers.AutoModelForDocumentQuestionAnswering[[transformers.AutoModelForDocumentQuestionAnswering]]

```python
transformers.AutoModelForDocumentQuestionAnswering(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2420)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 문서 질의응답 헤드를 포함하여 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForDocumentQuestionAnswering.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다: - [LayoutLMConfig](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMConfig) 구성 클래스: [LayoutLMForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMForQuestionAnswering) (LayoutLMConfig 모델) - [LayoutLMv2Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2Config) 구성 클래스: [LayoutLMv2ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2ForQuestionAnswering) (LayoutLMv2Config 모델) - [LayoutLMv3Config](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3Config) 구성 클래스: [LayoutLMv3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3ForQuestionAnswering) (LayoutLMv3Config 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForDocumentQuestionAnswering

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("impira/layoutlm-document-qa", revision="52e01b3")
>>> model = AutoModelForDocumentQuestionAnswering.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForDocumentQuestionAnswering.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **layoutlm** -- [LayoutLMForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/layoutlm#transformers.LayoutLMForQuestionAnswering) (LayoutLMConfig 모델)
- **layoutlmv2** -- [LayoutLMv2ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv2#transformers.LayoutLMv2ForQuestionAnswering) (LayoutLMv2Config 모델)
- **layoutlmv3** -- [LayoutLMv3ForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/layoutlmv3#transformers.LayoutLMv3ForQuestionAnswering) (LayoutLMv3Config 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForDocumentQuestionAnswering

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForDocumentQuestionAnswering.from_pretrained("impira/layoutlm-document-qa", revision="52e01b3")

>>> # Update configuration during loading
>>> model = AutoModelForDocumentQuestionAnswering.from_pretrained("impira/layoutlm-document-qa", revision="52e01b3", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForVisualQuestionAnswering[[transformers.AutoModelForVisualQuestionAnswering]]

#### transformers.AutoModelForVisualQuestionAnswering[[transformers.AutoModelForVisualQuestionAnswering]]

```python
transformers.AutoModelForVisualQuestionAnswering(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2409)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나(시각 질답 헤드를 포함)로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForVisualQuestionAnswering.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다: - [Blip2Config](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2Config) 구성 클래스: [Blip2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2ForConditionalGeneration) (Blip2Config 모델) - [BlipConfig](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipConfig) 구성 클래스: [BlipForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipForQuestionAnswering) (BlipConfig 모델) - [ViltConfig](/docs/transformers/v5.23.0/ko/model_doc/vilt#transformers.ViltConfig) 구성 클래스: [ViltForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/vilt#transformers.ViltForQuestionAnswering) (ViltConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForVisualQuestionAnswering

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("dandelin/vilt-b32-finetuned-vqa")
>>> model = AutoModelForVisualQuestionAnswering.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForVisualQuestionAnswering.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **blip** -- [BlipForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipForQuestionAnswering) (BlipConfig 모델)
- **blip-2** -- [Blip2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2ForConditionalGeneration) (Blip2Config 모델)
- **vilt** -- [ViltForQuestionAnswering](/docs/transformers/v5.23.0/ko/model_doc/vilt#transformers.ViltForQuestionAnswering) (ViltConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForVisualQuestionAnswering

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForVisualQuestionAnswering.from_pretrained("dandelin/vilt-b32-finetuned-vqa")

>>> # Update configuration during loading
>>> model = AutoModelForVisualQuestionAnswering.from_pretrained("dandelin/vilt-b32-finetuned-vqa", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForImageTextToText[[transformers.AutoModelForImageTextToText]]

#### transformers.AutoModelForImageTextToText[[transformers.AutoModelForImageTextToText]]

```python
transformers.AutoModelForImageTextToText(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2557)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다(이미지-텍스트-투-텍스트 모델링 헤드를 포함).

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForImageTextToText.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : The model class to instantiate is selected based on the configuration class:  - [AriaConfig](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaConfig) configuration class: [AriaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaForConditionalGeneration) (AriaConfig model) - [AyaVisionConfig](/docs/transformers/v5.23.0/ko/model_doc/aya_vision#transformers.AyaVisionConfig) configuration class: [AyaVisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/aya_vision#transformers.AyaVisionForConditionalGeneration) (AyaVisionConfig model) - [Blip2Config](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2Config) configuration class: [Blip2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2ForConditionalGeneration) (Blip2Config model) - [BlipConfig](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipConfig) configuration class: [BlipForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipForConditionalGeneration) (BlipConfig model) - [ChameleonConfig](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonConfig) configuration class: [ChameleonForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonForConditionalGeneration) (ChameleonConfig model) - [Cohere2VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere2_vision#transformers.Cohere2VisionConfig) configuration class: [Cohere2VisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cohere2_vision#transformers.Cohere2VisionForConditionalGeneration) (Cohere2VisionConfig model) - [CohereCompassConfig](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassConfig) configuration class: [CohereCompassForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassForConditionalGeneration) (CohereCompassConfig model) - [Cosmos3EdgeConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeConfig) configuration class: [Cosmos3EdgeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeForConditionalGeneration) (Cosmos3EdgeConfig model) - [Cosmos3OmniConfig](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_omni#transformers.Cosmos3OmniConfig) configuration class: [Cosmos3OmniForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_omni#transformers.Cosmos3OmniForConditionalGeneration) (Cosmos3OmniConfig model) - [DeepseekOcr2Config](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2Config) configuration class: [DeepseekOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2ForConditionalGeneration) (DeepseekOcr2Config model) - [DeepseekVLConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl#transformers.DeepseekVLConfig) configuration class: [DeepseekVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl#transformers.DeepseekVLForConditionalGeneration) (DeepseekVLConfig model) - [DeepseekVLHybridConfig](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl_hybrid#transformers.DeepseekVLHybridConfig) configuration class: [DeepseekVLHybridForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl_hybrid#transformers.DeepseekVLHybridForConditionalGeneration) (DeepseekVLHybridConfig model) - [DiffusionGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/diffusion_gemma#transformers.DiffusionGemmaConfig) configuration class: [DiffusionGemmaForBlockDiffusion](/docs/transformers/v5.23.0/ko/model_doc/diffusion_gemma#transformers.DiffusionGemmaForBlockDiffusion) (DiffusionGemmaConfig model) - [Dots3NoteConfig](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteConfig) configuration class: [Dots3NoteForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteForConditionalGeneration) (Dots3NoteConfig model) - [Emu3Config](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3Config) configuration class: [Emu3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3ForConditionalGeneration) (Emu3Config model) - [Ernie4_5_VLMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeConfig) configuration class: [Ernie4_5_VLMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeForConditionalGeneration) (Ernie4_5_VLMoeConfig model) - [EvollaConfig](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaConfig) configuration class: [EvollaForProteinText2Text](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaForProteinText2Text) (EvollaConfig model) - [Exaone4_5_Config](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_Config) configuration class: [Exaone4_5_ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_ForConditionalGeneration) (Exaone4_5_Config model) - [FastVlmConfig](/docs/transformers/v5.23.0/ko/model_doc/fast_vlm#transformers.FastVlmConfig) configuration class: [FastVlmForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fast_vlm#transformers.FastVlmForConditionalGeneration) (FastVlmConfig model) - [Florence2Config](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2Config) configuration class: [Florence2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2ForConditionalGeneration) (Florence2Config model) - [FuyuConfig](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuConfig) configuration class: [FuyuForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuForCausalLM) (FuyuConfig model) - [Gemma3Config](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3Config) configuration class: [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (Gemma3Config model) - [Gemma3nConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nConfig) configuration class: [Gemma3nForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nForConditionalGeneration) (Gemma3nConfig model) - [Gemma4Config](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4Config) configuration class: [Gemma4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4ForConditionalGeneration) (Gemma4Config model) - [Gemma4UnifiedConfig](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedConfig) configuration class: [Gemma4UnifiedForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedForConditionalGeneration) (Gemma4UnifiedConfig model) - [GitConfig](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitConfig) configuration class: [GitForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitForCausalLM) (GitConfig model) - [Glm46VConfig](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VConfig) configuration class: [Glm46VForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VForConditionalGeneration) (Glm46VConfig model) - [Glm4vConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vConfig) configuration class: [Glm4vForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vForConditionalGeneration) (Glm4vConfig model) - [Glm4vMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeConfig) configuration class: [Glm4vMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeForConditionalGeneration) (Glm4vMoeConfig model) - [Glm5NextConfig](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextConfig) configuration class: [Glm5NextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextForConditionalGeneration) (Glm5NextConfig model) - [GlmOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrConfig) configuration class: [GlmOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrForConditionalGeneration) (GlmOcrConfig model) - [GlmgaConfig](/docs/transformers/v5.23.0/ko/model_doc/glmga#transformers.GlmgaConfig) configuration class: [Glm46VForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VForConditionalGeneration) (GlmgaConfig model) - [GotOcr2Config](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2Config) configuration class: [GotOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2ForConditionalGeneration) (GotOcr2Config model) - [Granite4VisionConfig](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionConfig) configuration class: [Granite4VisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionForConditionalGeneration) (Granite4VisionConfig model) - [GraniteDoclingConfig](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingConfig) configuration class: [GraniteDoclingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingForConditionalGeneration) (GraniteDoclingConfig model) - [HunYuanVLConfig](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLConfig) configuration class: [HunYuanVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLForConditionalGeneration) (HunYuanVLConfig model) - [HyperCLOVAXVisionV2Config](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax_vision_v2#transformers.HyperCLOVAXVisionV2Config) configuration class: [HyperCLOVAXVisionV2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax_vision_v2#transformers.HyperCLOVAXVisionV2ForConditionalGeneration) (HyperCLOVAXVisionV2Config model) - [Idefics2Config](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2Config) configuration class: [Idefics2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2ForConditionalGeneration) (Idefics2Config model) - [Idefics3Config](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3Config) configuration class: [Idefics3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3ForConditionalGeneration) (Idefics3Config model) - [IdeficsConfig](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsConfig) configuration class: [IdeficsForVisionText2Text](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsForVisionText2Text) (IdeficsConfig model) - [Inkling2Config](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2Config) configuration class: [Inkling2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2ForConditionalGeneration) (Inkling2Config model) - [InklingConfig](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingConfig) configuration class: [InklingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingForConditionalGeneration) (InklingConfig model) - [InstructBlipConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipConfig) configuration class: [InstructBlipForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipForConditionalGeneration) (InstructBlipConfig model) - [InstructBlipVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoConfig) configuration class: [InstructBlipVideoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoForConditionalGeneration) (InstructBlipVideoConfig model) - [InternVLConfig](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLConfig) configuration class: [InternVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLForConditionalGeneration) (InternVLConfig model) - [JanusConfig](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusConfig) configuration class: [JanusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusForConditionalGeneration) (JanusConfig model) - [Kimi_K25Config](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25Config) configuration class: [Kimi_K25ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25ForConditionalGeneration) (Kimi_K25Config model) - [Kosmos2Config](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2Config) configuration class: [Kosmos2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2ForConditionalGeneration) (Kosmos2Config model) - [Kosmos2_5Config](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5Config) configuration class: [Kosmos2_5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5ForConditionalGeneration) (Kosmos2_5Config model) - [Lfm2VlConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm2_vl#transformers.Lfm2VlConfig) configuration class: [Lfm2VlForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lfm2_vl#transformers.Lfm2VlForConditionalGeneration) (Lfm2VlConfig model) - [Lfm3VlConfig](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlConfig) configuration class: [Lfm3VlForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlForConditionalGeneration) (Lfm3VlConfig model) - [LightOnOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/lighton_ocr#transformers.LightOnOcrConfig) configuration class: [LightOnOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lighton_ocr#transformers.LightOnOcrForConditionalGeneration) (LightOnOcrConfig model) - [Llama4Config](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4Config) configuration class: [Llama4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4ForConditionalGeneration) (Llama4Config model) - [LlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaConfig) configuration class: [LlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaForConditionalGeneration) (LlavaConfig model) - [LlavaNextConfig](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextConfig) configuration class: [LlavaNextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextForConditionalGeneration) (LlavaNextConfig model) - [LlavaNextVideoConfig](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoConfig) configuration class: [LlavaNextVideoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoForConditionalGeneration) (LlavaNextVideoConfig model) - [LlavaOnevisionConfig](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionConfig) configuration class: [LlavaOnevisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionForConditionalGeneration) (LlavaOnevisionConfig model) - [MiniCPMOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrConfig) configuration class: [MiniCPMOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrForConditionalGeneration) (MiniCPMOcrConfig model) - [MiniCPMV4_6Config](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6Config) configuration class: [MiniCPMV4_6ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6ForConditionalGeneration) (MiniCPMV4_6Config model) - [MiniCPMV4_7Config](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7Config) configuration class: [MiniCPMV4_7ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7ForConditionalGeneration) (MiniCPMV4_7Config model) - [MiniMaxM3VLConfig](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3VLConfig) configuration class: [MiniMaxM3SparseForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3SparseForConditionalGeneration) (MiniMaxM3VLConfig model) - [Mistral3Config](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3Config) configuration class: [Mistral3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3ForConditionalGeneration) (Mistral3Config model) - [Mistral4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4Config) configuration class: [Mistral4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4ForCausalLM) (Mistral4Config model) - [MistralLarge4Config](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4Config) configuration class: [MistralLarge4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4ForConditionalGeneration) (MistralLarge4Config model) - [MllamaConfig](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaConfig) configuration class: [MllamaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaForConditionalGeneration) (MllamaConfig model) - [Molmo2Config](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2Config) configuration class: [Molmo2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2ForConditionalGeneration) (Molmo2Config model) - [MuseGlimmerConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerConfig) configuration class: [MuseGlimmerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerForConditionalGeneration) (MuseGlimmerConfig model) - [MuseSparkConfig](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkConfig) configuration class: [MuseSparkForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkForConditionalGeneration) (MuseSparkConfig model) - [NemotronH_Omni_Reasoning_V3_Config](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h_omni#transformers.NemotronH_Omni_Reasoning_V3_Config) configuration class: [NemotronH_Omni_Reasoning_V3](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h_omni#transformers.NemotronH_Omni_Reasoning_V3) (NemotronH_Omni_Reasoning_V3_Config model) - [Ovis2Config](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2Config) configuration class: [Ovis2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2ForConditionalGeneration) (Ovis2Config model) - [PI0Config](/docs/transformers/v5.23.0/ko/model_doc/pi0#transformers.PI0Config) configuration class: [PI0ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pi0#transformers.PI0ForConditionalGeneration) (PI0Config model) - [PPChart2TableConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_chart2table#transformers.PPChart2TableConfig) configuration class: [GotOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2ForConditionalGeneration) (PPChart2TableConfig model) - [PPFormulaNetConfig](/docs/transformers/v5.23.0/ko/model_doc/pp_formulanet#transformers.PPFormulaNetConfig) configuration class: [PPFormulaNetForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pp_formulanet#transformers.PPFormulaNetForConditionalGeneration) (PPFormulaNetConfig model) - [PaddleOCRVLConfig](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVLConfig) configuration class: [PaddleOCRVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVLForConditionalGeneration) (PaddleOCRVLConfig model) - [PaliGemmaConfig](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaConfig) configuration class: [PaliGemmaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaForConditionalGeneration) (PaliGemmaConfig model) - [PerceptionLMConfig](/docs/transformers/v5.23.0/ko/model_doc/perception_lm#transformers.PerceptionLMConfig) configuration class: [PerceptionLMForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/perception_lm#transformers.PerceptionLMForConditionalGeneration) (PerceptionLMConfig model) - [Pix2StructConfig](/docs/transformers/v5.23.0/ko/model_doc/pix2struct#transformers.Pix2StructConfig) configuration class: [Pix2StructForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pix2struct#transformers.Pix2StructForConditionalGeneration) (Pix2StructConfig model) - [QianfanOCRConfig](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRConfig) configuration class: [QianfanOCRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRForConditionalGeneration) (QianfanOCRConfig model) - [Qwen2VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLConfig) configuration class: [Qwen2VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLForConditionalGeneration) (Qwen2VLConfig model) - [Qwen2_5OmniThinkerConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniThinkerConfig) configuration class: [Qwen2_5OmniThinkerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniThinkerForConditionalGeneration) (Qwen2_5OmniThinkerConfig model) - [Qwen2_5_VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLConfig) configuration class: [Qwen2_5_VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLForConditionalGeneration) (Qwen2_5_VLConfig model) - [Qwen3OmniMoeThinkerConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeThinkerConfig) configuration class: [Qwen3OmniMoeThinkerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeThinkerForConditionalGeneration) (Qwen3OmniMoeThinkerConfig model) - [Qwen3VLConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLConfig) configuration class: [Qwen3VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLForConditionalGeneration) (Qwen3VLConfig model) - [Qwen3VLMoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeConfig) configuration class: [Qwen3VLMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeForConditionalGeneration) (Qwen3VLMoeConfig model) - [Qwen3_5Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Config) configuration class: [Qwen3_5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForConditionalGeneration) (Qwen3_5Config model) - [Qwen3_5MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeConfig) configuration class: [Qwen3_5MoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeForConditionalGeneration) (Qwen3_5MoeConfig model) - [Qwen4ExpConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpConfig) configuration class: [Qwen4ExpForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpForConditionalGeneration) (Qwen4ExpConfig model) - [ShieldGemma2Config](/docs/transformers/v5.23.0/ko/model_doc/shieldgemma2#transformers.ShieldGemma2Config) configuration class: [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (ShieldGemma2Config model) - [SmolVLMConfig](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMConfig) configuration class: [SmolVLMForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMForConditionalGeneration) (SmolVLMConfig model) - [Step3p7Config](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7Config) configuration class: [Step3p7ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7ForConditionalGeneration) (Step3p7Config model) - [T5Gemma2Config](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2Config) configuration class: [T5Gemma2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForConditionalGeneration) (T5Gemma2Config model) - [UdopConfig](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopConfig) configuration class: [UdopForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopForConditionalGeneration) (UdopConfig model) - [UnlimitedOcrConfig](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrConfig) configuration class: [UnlimitedOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrForConditionalGeneration) (UnlimitedOcrConfig model) - [VideoLlama3Config](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3Config) configuration class: [VideoLlama3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3ForConditionalGeneration) (VideoLlama3Config model) - [VideoLlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaConfig) configuration class: [VideoLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaForConditionalGeneration) (VideoLlavaConfig model) - [VipLlavaConfig](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaConfig) configuration class: [VipLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaForConditionalGeneration) (VipLlavaConfig model) - [VisionEncoderDecoderConfig](/docs/transformers/v5.23.0/ko/model_doc/vision-encoder-decoder#transformers.VisionEncoderDecoderConfig) configuration class: [VisionEncoderDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/vision-encoder-decoder#transformers.VisionEncoderDecoderModel) (VisionEncoderDecoderConfig model)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForImageTextToText

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForImageTextToText.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForImageTextToText.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **aria** -- [AriaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/aria#transformers.AriaForConditionalGeneration) (AriaConfig 모델)
- **aya_vision** -- [AyaVisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/aya_vision#transformers.AyaVisionForConditionalGeneration) (AyaVisionConfig 모델)
- **blip** -- [BlipForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blip#transformers.BlipForConditionalGeneration) (BlipConfig 모델)
- **blip-2** -- [Blip2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/blip-2#transformers.Blip2ForConditionalGeneration) (Blip2Config 모델)
- **chameleon** -- [ChameleonForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/chameleon#transformers.ChameleonForConditionalGeneration) (ChameleonConfig 모델)
- **cohere2_vision** -- [Cohere2VisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cohere2_vision#transformers.Cohere2VisionForConditionalGeneration) (Cohere2VisionConfig 모델)
- **cohere_compass** -- [CohereCompassForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cohere_compass#transformers.CohereCompassForConditionalGeneration) (CohereCompassConfig 모델)
- **cosmos3_edge** -- [Cosmos3EdgeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_edge#transformers.Cosmos3EdgeForConditionalGeneration) (Cosmos3EdgeConfig 모델)
- **cosmos3_omni** -- [Cosmos3OmniForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/cosmos3_omni#transformers.Cosmos3OmniForConditionalGeneration) (Cosmos3OmniConfig 모델)
- **deepseek_ocr2** -- [DeepseekOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_ocr2#transformers.DeepseekOcr2ForConditionalGeneration) (DeepseekOcr2Config 모델)
- **deepseek_vl** -- [DeepseekVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl#transformers.DeepseekVLForConditionalGeneration) (DeepseekVLConfig 모델)
- **deepseek_vl_hybrid** -- [DeepseekVLHybridForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/deepseek_vl_hybrid#transformers.DeepseekVLHybridForConditionalGeneration) (DeepseekVLHybridConfig 모델)
- **diffusion_gemma** -- [DiffusionGemmaForBlockDiffusion](/docs/transformers/v5.23.0/ko/model_doc/diffusion_gemma#transformers.DiffusionGemmaForBlockDiffusion) (DiffusionGemmaConfig 모델)
- **dots3_note** -- [Dots3NoteForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/dots3_note#transformers.Dots3NoteForConditionalGeneration) (Dots3NoteConfig 모델)
- **emu3** -- [Emu3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/emu3#transformers.Emu3ForConditionalGeneration) (Emu3Config 모델)
- **ernie4_5_vl_moe** -- [Ernie4_5_VLMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/ernie4_5_vl_moe#transformers.Ernie4_5_VLMoeForConditionalGeneration) (Ernie4_5_VLMoeConfig 모델)
- **evolla** -- [EvollaForProteinText2Text](/docs/transformers/v5.23.0/ko/model_doc/evolla#transformers.EvollaForProteinText2Text) (EvollaConfig 모델)
- **exaone4_5** -- [Exaone4_5_ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/exaone4_5#transformers.Exaone4_5_ForConditionalGeneration) (Exaone4_5_Config 모델)
- **fast_vlm** -- [FastVlmForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/fast_vlm#transformers.FastVlmForConditionalGeneration) (FastVlmConfig 모델)
- **florence2** -- [Florence2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/florence2#transformers.Florence2ForConditionalGeneration) (Florence2Config 모델)
- **fuyu** -- [FuyuForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/fuyu#transformers.FuyuForCausalLM) (FuyuConfig 모델)
- **gemma3** -- [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (Gemma3Config 모델)
- **gemma3n** -- [Gemma3nForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3n#transformers.Gemma3nForConditionalGeneration) (Gemma3nConfig 모델)
- **gemma4** -- [Gemma4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4ForConditionalGeneration) (Gemma4Config 모델)
- **gemma4_unified** -- [Gemma4UnifiedForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma4_unified#transformers.Gemma4UnifiedForConditionalGeneration) (Gemma4UnifiedConfig 모델)
- **git** -- [GitForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/git#transformers.GitForCausalLM) (GitConfig 모델)
- **glm46v** -- [Glm46VForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VForConditionalGeneration) (Glm46VConfig 모델)
- **glm4v** -- [Glm4vForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm4v#transformers.Glm4vForConditionalGeneration) (Glm4vConfig 모델)
- **glm4v_moe** -- [Glm4vMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm4v_moe#transformers.Glm4vMoeForConditionalGeneration) (Glm4vMoeConfig 모델)
- **glm5_next** -- [Glm5NextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm5_next#transformers.Glm5NextForConditionalGeneration) (Glm5NextConfig 모델)
- **glm_ocr** -- [GlmOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm_ocr#transformers.GlmOcrForConditionalGeneration) (GlmOcrConfig 모델)
- **glmga** -- [Glm46VForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/glm46v#transformers.Glm46VForConditionalGeneration) (GlmgaConfig 모델)
- **got_ocr2** -- [GotOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2ForConditionalGeneration) (GotOcr2Config 모델)
- **granite4_vision** -- [Granite4VisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite4_vision#transformers.Granite4VisionForConditionalGeneration) (Granite4VisionConfig 모델)
- **granite_docling** -- [GraniteDoclingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granite_docling#transformers.GraniteDoclingForConditionalGeneration) (GraniteDoclingConfig 모델)
- **hunyuan_vl** -- [HunYuanVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/hunyuan_vl#transformers.HunYuanVLForConditionalGeneration) (HunYuanVLConfig 모델)
- **hyperclovax_vision_v2** -- [HyperCLOVAXVisionV2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/hyperclovax_vision_v2#transformers.HyperCLOVAXVisionV2ForConditionalGeneration) (HyperCLOVAXVisionV2Config 모델)
- **idefics** -- [IdeficsForVisionText2Text](/docs/transformers/v5.23.0/ko/model_doc/idefics#transformers.IdeficsForVisionText2Text) (IdeficsConfig 모델)
- **idefics2** -- [Idefics2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics2#transformers.Idefics2ForConditionalGeneration) (Idefics2Config 모델)
- **idefics3** -- [Idefics3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/idefics3#transformers.Idefics3ForConditionalGeneration) (Idefics3Config 모델)
- **inkling2** -- [Inkling2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling2#transformers.Inkling2ForConditionalGeneration) (Inkling2Config 모델)
- **inkling_mm_model** -- [InklingForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/inkling#transformers.InklingForConditionalGeneration) (InklingConfig 모델)
- **instructblip** -- [InstructBlipForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/instructblip#transformers.InstructBlipForConditionalGeneration) (InstructBlipConfig 모델)
- **instructblipvideo** -- [InstructBlipVideoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/instructblipvideo#transformers.InstructBlipVideoForConditionalGeneration) (InstructBlipVideoConfig 모델)
- **internvl** -- [InternVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/internvl#transformers.InternVLForConditionalGeneration) (InternVLConfig 모델)
- **janus** -- [JanusForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/janus#transformers.JanusForConditionalGeneration) (JanusConfig 모델)
- **kimi_k25** -- [Kimi_K25ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kimi_k25#transformers.Kimi_K25ForConditionalGeneration) (Kimi_K25Config 모델)
- **kosmos-2** -- [Kosmos2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kosmos-2#transformers.Kosmos2ForConditionalGeneration) (Kosmos2Config 모델)
- **kosmos-2.5** -- [Kosmos2_5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/kosmos2_5#transformers.Kosmos2_5ForConditionalGeneration) (Kosmos2_5Config 모델)
- **lfm2_vl** -- [Lfm2VlForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lfm2_vl#transformers.Lfm2VlForConditionalGeneration) (Lfm2VlConfig 모델)
- **lfm3_vl** -- [Lfm3VlForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lfm3_vl#transformers.Lfm3VlForConditionalGeneration) (Lfm3VlConfig 모델)
- **lighton_ocr** -- [LightOnOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/lighton_ocr#transformers.LightOnOcrForConditionalGeneration) (LightOnOcrConfig 모델)
- **llama4** -- [Llama4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llama4#transformers.Llama4ForConditionalGeneration) (Llama4Config 모델)
- **llava** -- [LlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava#transformers.LlavaForConditionalGeneration) (LlavaConfig 모델)
- **llava_next** -- [LlavaNextForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/granitevision#transformers.LlavaNextForConditionalGeneration) (LlavaNextConfig 모델)
- **llava_next_video** -- [LlavaNextVideoForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_next_video#transformers.LlavaNextVideoForConditionalGeneration) (LlavaNextVideoConfig 모델)
- **llava_onevision** -- [LlavaOnevisionForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/llava_onevision#transformers.LlavaOnevisionForConditionalGeneration) (LlavaOnevisionConfig 모델)
- **minicpm_ocr** -- [MiniCPMOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpm_ocr#transformers.MiniCPMOcrForConditionalGeneration) (MiniCPMOcrConfig 모델)
- **minicpmv4_6** -- [MiniCPMV4_6ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_6#transformers.MiniCPMV4_6ForConditionalGeneration) (MiniCPMV4_6Config 모델)
- **minicpmv4_7** -- [MiniCPMV4_7ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minicpmv4_7#transformers.MiniCPMV4_7ForConditionalGeneration) (MiniCPMV4_7Config 모델)
- **minimax_m3_vl** -- [MiniMaxM3SparseForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/minimax_m3_vl#transformers.MiniMaxM3SparseForConditionalGeneration) (MiniMaxM3VLConfig 모델)
- **mistral3** -- [Mistral3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral3#transformers.Mistral3ForConditionalGeneration) (Mistral3Config 모델)
- **mistral4** -- [Mistral4ForCausalLM](/docs/transformers/v5.23.0/ko/model_doc/mistral4#transformers.Mistral4ForCausalLM) (Mistral4Config 모델)
- **mistral_large4** -- [MistralLarge4ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mistral_large4#transformers.MistralLarge4ForConditionalGeneration) (MistralLarge4Config 모델)
- **mllama** -- [MllamaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/mllama#transformers.MllamaForConditionalGeneration) (MllamaConfig 모델)
- **molmo2** -- [Molmo2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/molmo2#transformers.Molmo2ForConditionalGeneration) (Molmo2Config 모델)
- **muse_glimmer** -- [MuseGlimmerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/muse_glimmer#transformers.MuseGlimmerForConditionalGeneration) (MuseGlimmerConfig 모델)
- **muse_spark** -- [MuseSparkForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/muse_spark#transformers.MuseSparkForConditionalGeneration) (MuseSparkConfig 모델)
- **nemotron_h_omni** -- [NemotronH_Omni_Reasoning_V3](/docs/transformers/v5.23.0/ko/model_doc/nemotron_h_omni#transformers.NemotronH_Omni_Reasoning_V3) (NemotronH_Omni_Reasoning_V3_Config 모델)
- **ovis2** -- [Ovis2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/ovis2#transformers.Ovis2ForConditionalGeneration) (Ovis2Config 모델)
- **paddleocr_vl** -- [PaddleOCRVLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/paddleocr_vl#transformers.PaddleOCRVLForConditionalGeneration) (PaddleOCRVLConfig 모델)
- **paligemma** -- [PaliGemmaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/paligemma#transformers.PaliGemmaForConditionalGeneration) (PaliGemmaConfig 모델)
- **perception_lm** -- [PerceptionLMForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/perception_lm#transformers.PerceptionLMForConditionalGeneration) (PerceptionLMConfig 모델)
- **pi0** -- [PI0ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pi0#transformers.PI0ForConditionalGeneration) (PI0Config 모델)
- **pix2struct** -- [Pix2StructForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pix2struct#transformers.Pix2StructForConditionalGeneration) (Pix2StructConfig 모델)
- **pp_chart2table** -- [GotOcr2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/got_ocr2#transformers.GotOcr2ForConditionalGeneration) (PPChart2TableConfig 모델)
- **pp_formulanet** -- [PPFormulaNetForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/pp_formulanet#transformers.PPFormulaNetForConditionalGeneration) (PPFormulaNetConfig 모델)
- **qianfan_ocr** -- [QianfanOCRForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qianfan_ocr#transformers.QianfanOCRForConditionalGeneration) (QianfanOCRConfig 모델)
- **qwen2_5_omni_thinker** -- [Qwen2_5OmniThinkerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_omni#transformers.Qwen2_5OmniThinkerForConditionalGeneration) (Qwen2_5OmniThinkerConfig 모델)
- **qwen2_5_vl** -- [Qwen2_5_VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_5_vl#transformers.Qwen2_5_VLForConditionalGeneration) (Qwen2_5_VLConfig 모델)
- **qwen2_vl** -- [Qwen2VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen2_vl#transformers.Qwen2VLForConditionalGeneration) (Qwen2VLConfig 모델)
- **qwen3_5** -- [Qwen3_5ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5ForConditionalGeneration) (Qwen3_5Config 모델)
- **qwen3_5_moe** -- [Qwen3_5MoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeForConditionalGeneration) (Qwen3_5MoeConfig 모델)
- **qwen3_omni_moe_thinker** -- [Qwen3OmniMoeThinkerForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_omni_moe#transformers.Qwen3OmniMoeThinkerForConditionalGeneration) (Qwen3OmniMoeThinkerConfig 모델)
- **qwen3_vl** -- [Qwen3VLForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl#transformers.Qwen3VLForConditionalGeneration) (Qwen3VLConfig 모델)
- **qwen3_vl_moe** -- [Qwen3VLMoeForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen3_vl_moe#transformers.Qwen3VLMoeForConditionalGeneration) (Qwen3VLMoeConfig 모델)
- **qwen4_exp** -- [Qwen4ExpForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/qwen4_exp#transformers.Qwen4ExpForConditionalGeneration) (Qwen4ExpConfig 모델)
- **shieldgemma2** -- [Gemma3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/gemma3#transformers.Gemma3ForConditionalGeneration) (ShieldGemma2Config 모델)
- **smolvlm** -- [SmolVLMForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/smolvlm#transformers.SmolVLMForConditionalGeneration) (SmolVLMConfig 모델)
- **step3p7** -- [Step3p7ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/step3p7#transformers.Step3p7ForConditionalGeneration) (Step3p7Config 모델)
- **t5gemma2** -- [T5Gemma2ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/t5gemma2#transformers.T5Gemma2ForConditionalGeneration) (T5Gemma2Config 모델)
- **udop** -- [UdopForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/udop#transformers.UdopForConditionalGeneration) (UdopConfig 모델)
- **unlimited_ocr** -- [UnlimitedOcrForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/unlimited_ocr#transformers.UnlimitedOcrForConditionalGeneration) (UnlimitedOcrConfig 모델)
- **video_llama_3** -- [VideoLlama3ForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/video_llama_3#transformers.VideoLlama3ForConditionalGeneration) (VideoLlama3Config 모델)
- **video_llava** -- [VideoLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/video_llava#transformers.VideoLlavaForConditionalGeneration) (VideoLlavaConfig 모델)
- **vipllava** -- [VipLlavaForConditionalGeneration](/docs/transformers/v5.23.0/ko/model_doc/vipllava#transformers.VipLlavaForConditionalGeneration) (VipLlavaConfig 모델)
- **vision-encoder-decoder** -- [VisionEncoderDecoderModel](/docs/transformers/v5.23.0/ko/model_doc/vision-encoder-decoder#transformers.VisionEncoderDecoderModel) (VisionEncoderDecoderConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForImageTextToText

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForImageTextToText.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForImageTextToText.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

### AutoModelForDecision[[transformers.AutoModelForDecision]]

#### transformers.AutoModelForDecision[[transformers.AutoModelForDecision]]

```python
transformers.AutoModelForDecision(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2662)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성될 때, 라이브러리의 모델 클래스 중 하나로(결정 헤드를 포함하여) 인스턴스화되는 범용 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForDecision.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 구성 클래스에 따라 선택됩니다: - [Gemma4Config](/docs/transformers/v5.23.0/ko/model_doc/gemma4#transformers.Gemma4Config) 구성 클래스: [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (Gemma4Config 모델) - [JevConfig](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevConfig) 구성 클래스: [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (JevConfig 모델) - [Qwen2Config](/docs/transformers/v5.23.0/ko/model_doc/qwen2#transformers.Qwen2Config) 구성 클래스: [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (Qwen2Config 모델) - [Qwen3_5Config](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5Config) 구성 클래스: [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (Qwen3_5Config 모델) - [Qwen3_5MoeConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5_moe#transformers.Qwen3_5MoeConfig) 구성 클래스: [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (Qwen3_5MoeConfig 모델) - [Qwen3_5TextConfig](/docs/transformers/v5.23.0/ko/model_doc/qwen3_5#transformers.Qwen3_5TextConfig) 구성 클래스: [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (Qwen3_5TextConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForDecision

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForDecision.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForDecision.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **gemma4** -- [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (Gemma4Config 모델)
- **jev** -- [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (JevConfig 모델)
- **qwen2** -- [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (Qwen2Config 모델)
- **qwen3_5** -- [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (Qwen3_5Config 모델)
- **qwen3_5_moe** -- [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (Qwen3_5MoeConfig 모델)
- **qwen3_5_text** -- [JevForDecision](/docs/transformers/v5.23.0/ko/model_doc/jev#transformers.JevForDecision) (Qwen3_5TextConfig 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForDecision

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForDecision.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForDecision.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

## 시계열

### AutoModelForTimeSeriesPrediction[[transformers.AutoModelForTimeSeriesPrediction]]

#### transformers.AutoModelForTimeSeriesPrediction[[transformers.AutoModelForTimeSeriesPrediction]]

```python
transformers.AutoModelForTimeSeriesPrediction(*args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/modeling_auto.py#L2486)

이것은 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained) 클래스 메서드 또는 [from_config()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_config) 클래스 메서드로 생성되면, 시계열 예측 헤드를 갖춘 라이브러리의 모델 클래스 중 하나로 인스턴스화되는 일반적인 모델 클래스입니다.

이 클래스는 `__init__()`를 사용하여 직접 인스턴스화할 수 없습니다(에러가 발생합니다).

#### from_config[[transformers.AutoModelForTimeSeriesPrediction.from_config]]

```python
from_config(**kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L204)

**매개변수:**

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig)) : 인스턴스화할 모델 클래스는 설정 클래스에 따라 선택됩니다:  - [TimesFm2_5Config](/docs/transformers/v5.23.0/ko/model_doc/timesfm2_5#transformers.TimesFm2_5Config) 설정 클래스: [TimesFm2_5ModelForPrediction](/docs/transformers/v5.23.0/ko/model_doc/timesfm2_5#transformers.TimesFm2_5ModelForPrediction) (TimesFm2_5Config 모델) - [TimesFmConfig](/docs/transformers/v5.23.0/ko/model_doc/timesfm#transformers.TimesFmConfig) 설정 클래스: [TimesFmModelForPrediction](/docs/transformers/v5.23.0/ko/model_doc/timesfm#transformers.TimesFmModelForPrediction) (TimesFmConfig 모델)

attn_implementation (`str`, *선택 사항*) : 모델에서 사용할 attention 구현 방식(해당하는 경우)입니다. 다음 중 하나를 사용할 수 있습니다. - `"eager"` (attention의 수동 구현) - `"sdpa"` ([`F.scaled_dot_product_attention`](https://pytorch.org/docs/master/generated/torch.nn.functional.scaled_dot_product_attention.html) 사용) - `"flash_attention_2"` ([Dao-AILab/flash-attention](https://github.com/Dao-AILab/flash-attention) 사용) - `"flash_attention_3"` ([Dao-AILab/flash-attention/hopper](https://github.com/Dao-AILab/flash-attention/tree/main/hopper) 사용) - `"flash_attention_4"` ([Dao-AILab/flash-attention/flash_attn/cute](https://github.com/Dao-AILab/flash-attention/tree/main/flash_attn/cute) 사용). 기본적으로 사용 가능한 경우, torch>=2.1.1에서는 SDPA가 사용됩니다. 그렇지 않으면 수동 `"eager"` 구현이 기본값입니다.

…

참고: 구성 파일에서 모델을 로드하는 것은 모델 가중치를 **로드하지 않습니다**. 모델 구성에만 영향을 미칩니다. 모델 가중치를 로드하려면 [from_pretrained()](/docs/transformers/v5.23.0/ko/model_doc/auto#transformers.AutoModel.from_pretrained)를 사용하세요.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTimeSeriesPrediction

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-cased")
>>> model = AutoModelForTimeSeriesPrediction.from_config(config)
```

#### from_pretrained[[transformers.AutoModelForTimeSeriesPrediction.from_pretrained]]

```python
from_pretrained(*model_args, **kwargs)
```

[소스](https://github.com/huggingface/transformers/blob/v5.23.0/src/transformers/models/auto/auto_factory.py#L260)

**매개변수:**

pretrained_model_name_or_path (`str` 또는 `os.PathLike`) : 다음 중 하나일 수 있습니다. - huggingface.co의 모델 저장소에 호스팅된 사전 학습 모델의 *model id*인 문자열. - [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 모델 가중치를 저장한 *directory*의 경로, 예: `./my_model_directory/`.

model_args (추가 위치 인자, *선택 사항*) : 기본 모델의 `__init__()` 메서드에 전달됩니다.

config ([PreTrainedConfig](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig), *선택*) : 자동으로 로드되는 설정 대신 사용할 모델의 설정입니다. 다음 경우 설정이 자동으로 로드됩니다:
- 모델이 라이브러리가 제공하는 모델인 경우 (학습된 모델의 *model id* 문자열로 로드됨).
- 모델이 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)를 사용하여 저장된 후 저장 디렉터리를 지정하여 다시 로드된 경우.
- 로컬 디렉터리를 `pretrained_model_name_or_path`로 지정하여 모델을 로드하고, 해당 디렉터리에 *config.json*이라는 이름의 설정 JSON 파일이 있는 경우.

state_dict (*dict[str, torch.Tensor]*, *선택*) : 저장된 가중치 파일에서 로드한 state dictionary 대신 사용할 state dictionary입니다. 사전 학습된 구성으로 모델을 생성하되 자체 가중치를 로드하려는 경우 이 옵션을 사용할 수 있습니다. 이 경우 [save_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.save_pretrained)와 [from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/model#transformers.PreTrainedModel.from_pretrained)를 사용하는 것이 더 간단한 옵션인지 확인해 보시기 바랍니다.

cache_dir (`str` 또는 `os.PathLike`, *선택 사항*) : 표준 캐시를 사용하지 않을 경우, 다운로드된 사전 학습된 모델 구성을 캐시할 디렉터리의 경로입니다.

force_download (`bool`, *선택 사항*, 기본값 `False`) : 캐시된 버전이 존재하더라도 모델 가중치 및 구성 파일의 (재)다운로드를 강제할지 여부입니다.

proxies (`dict[str, str]`, *선택*) : 프로토콜 또는 엔드포인트별로 사용할 프록시 서버의 사전입니다. 예: `{'http': 'foo.bar:3128', 'http://hostname': 'foo.bar:4012'}`. 각 요청에 프록시가 사용됩니다.

output_loading_info(`bool`, *선택 사항*, 기본값 `False`) : 누락된 키, 예상치 못한 키, 오류 메시지가 포함된 사전도 함께 반환할지 여부.

local_files_only(`bool`, *선택 사항*, 기본값 `False`) : 로컬 파일만 사용할지 여부(예: 모델 다운로드를 시도하지 않음).

revision (`str`, *선택 사항*, 기본값 `"main"`) : 사용할 특정 모델 버전입니다. huggingface.co에서 모델 및 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 브랜치 이름, 태그 이름 또는 커밋 ID가 될 수 있습니다.

trust_remote_code (`bool`, *선택 사항*, 기본값 `False`) : Hub에서 자체 modeling 파일로 정의된 커스텀 모델을 허용할지 여부입니다. 이 옵션은 Hub에 존재하는 코드가 로컬 머신에서 실행되므로, 신뢰할 수 있는 저장소에서만 `True`로 설정하고 해당 코드를 확인한 후에 사용해야 합니다.

code_revision (`str`, *선택 사항*, 기본값 `"main"`) : 모델의 나머지 부분과 다른 저장소에 코드가 있는 경우, Hub에서 사용할 특정 revision입니다. huggingface.co에서 모델과 기타 아티팩트를 저장하기 위해 git 기반 시스템을 사용하므로, `revision`은 git에서 허용하는 모든 식별자, 즉 branch 이름, tag 이름 또는 commit id가 될 수 있습니다.

kwargs (추가 키워드 인자, *선택 사항*) : 설정 객체를 로드한 후 업데이트하고 모델을 초기화하는 데 사용할 수 있습니다(예: `output_attentions=True`). `config`가 제공되었는지 자동으로 로드되었는지에 따라 동작이 다릅니다.  - `config`로 설정이 제공된 경우, `**kwargs`는 기본 모델의 `__init__` 메서드에 직접 전달됩니다(설정 관련 모든 업데이트가 이미 수행된 것으로 가정합니다). - 설정이 제공되지 않은 경우, `kwargs`는 먼저 설정 클래스 초기화 함수([from_pretrained()](/docs/transformers/v5.23.0/ko/main_classes/configuration#transformers.PreTrainedConfig.from_pretrained))에 전달됩니다. `kwargs`의 각 키 중 설정 속성에 해당하는 키는 제공된 `kwargs` 값으로 해당 속성을 덮어쓰는 데 사용됩니다. 설정 속성에 해당하지 않는 나머지 키는 기본 모델의 `__init__` 함수에 전달됩니다.

…

인스턴스화할 모델 클래스는 config 객체의 `model_type` 속성(인자로 전달되거나, 가능하면 `pretrained_model_name_or_path`에서 로드된 것)에 따라 선택됩니다. 해당 속성이 없으면 `pretrained_model_name_or_path`에 대해 패턴 매칭을 사용하는 방식으로 폴백합니다.

- **timesfm** -- [TimesFmModelForPrediction](/docs/transformers/v5.23.0/ko/model_doc/timesfm#transformers.TimesFmModelForPrediction) (TimesFmConfig 모델)
- **timesfm2_5** -- [TimesFm2_5ModelForPrediction](/docs/transformers/v5.23.0/ko/model_doc/timesfm2_5#transformers.TimesFm2_5ModelForPrediction) (TimesFm2_5Config 모델)

모델은 기본적으로 `model.eval()`를 사용하여 평가 모드로 설정됩니다(예를 들어, dropout 모듈이 비활성화됨). 모델을 학습하려면 먼저 `model.train()`를 사용하여 학습 모드로 되돌려야 합니다.

예:

```python
>>> from transformers import AutoConfig, AutoModelForTimeSeriesPrediction

>>> # Download model and configuration from huggingface.co and cache.
>>> model = AutoModelForTimeSeriesPrediction.from_pretrained("google-bert/bert-base-cased")

>>> # Update configuration during loading
>>> model = AutoModelForTimeSeriesPrediction.from_pretrained("google-bert/bert-base-cased", output_attentions=True)
>>> model.config.output_attentions
True
```

