SentenceTransformer based on BAAI/bge-base-en-v1.5
This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5 on the json dataset. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
Model Details
Model Description
- Model Type: Sentence Transformer
- Base model: BAAI/bge-base-en-v1.5
- Maximum Sequence Length: 512 tokens
- Output Dimensionality: 768 dimensions
- Similarity Function: Cosine Similarity
- Training Dataset:
- json
Model Sources
- Documentation: Sentence Transformers Documentation
- Repository: Sentence Transformers on GitHub
- Hugging Face: Sentence Transformers on Hugging Face
Full Model Architecture
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
(2): Normalize()
)
Usage
Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
pip install -U sentence-transformers
Then you can load this model and run inference.
from sentence_transformers import SentenceTransformer
# Download from the ๐ค Hub
model = SentenceTransformer("model")
# Run inference
sentences = [
'The weather is lovely today.',
"It's so sunny outside!",
'He drove to the stadium.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
Evaluation
Metrics
Semantic Similarity
- Dataset:
EmbeddingSimEval - Evaluated with
EmbeddingSimilarityEvaluator
| Metric | Value |
|---|---|
| pearson_cosine | nan |
| spearman_cosine | nan |
Binary Classification
- Dataset:
BinaryClassifEval - Evaluated with
BinaryClassificationEvaluator
| Metric | Value |
|---|---|
| cosine_accuracy | 0.8 |
| cosine_accuracy_threshold | 0.6527 |
| cosine_f1 | 0.8889 |
| cosine_f1_threshold | 0.6527 |
| cosine_precision | 1.0 |
| cosine_recall | 0.8 |
| cosine_ap | 1.0 |
| cosine_mcc | 0.0 |
Training Details
Training Dataset
json
- Dataset: json
- Size: 3,696 training samples
- Columns:
sentence1,sentence2, andlabel - Approximate statistics based on the first 1000 samples:
sentence1 sentence2 label type string string int details - min: 37 tokens
- mean: 40.4 tokens
- max: 44 tokens
- min: 49 tokens
- mean: 62.2 tokens
- max: 85 tokens
- 1: 100.00%
- Samples:
sentence1 sentence2 label Quelle proportion des dรฉpenses pour l'installation de bornes de recharge รฉlectrique peut รชtre couverte par la rรฉgion รle-de-France?Nature de l'aide: La Rรฉgion participera ร hauteur de 50% de la dรฉpense supportรฉe par le maรฎtre dโouvrage plafonnรฉe en fonction du type de bornes1Quels types de projets sont รฉligibles pour obtenir un financement de la Rรฉgion รle-de-France dans le cadre du dรฉveloppement de l'รฉlectromobilitรฉ?Type de project: Le dispositif a pour objet le financement : Des รฉtudes dโรฉlaboration dโun document stratรฉgique,De lโinstallation ou la mise ร niveau des IRVE situรฉes sur le domaine public francilien, respectant les critรจres du label rรฉgional et sโinscrivant dans un plan dโactions1Quelle est la dรฉmarche ร suivre pour dรฉposer une demande de subvention concernant l'รฉlectromobilitรฉ en รle-de-France?Procรฉdures et dรฉmarches: Dรฉposez sur mesdemarches.iledefrance.fr votre dossier de demande de subvention prรฉsentant le projet de maniรจre prรฉcise et comportant toutes les piรจces permettant lโinstruction du dossier, rรฉputรฉ complet, par les services de la Rรฉgion1 - Loss:
MultipleNegativesRankingLosswith these parameters:{ "scale": 20.0, "similarity_fct": "cos_sim" }
Evaluation Dataset
json
- Dataset: json
- Size: 687 evaluation samples
- Columns:
sentence1,sentence2, andlabel - Approximate statistics based on the first 687 samples:
sentence1 sentence2 label type string string int details - min: 24 tokens
- mean: 33.6 tokens
- max: 42 tokens
- min: 37 tokens
- mean: 90.4 tokens
- max: 257 tokens
- 1: 100.00%
- Samples:
sentence1 sentence2 label Sous quelles conditions mon centre de formation en apprentissage peut-il รชtre รฉligible ร une subvention pour des investissements?Le dispositif est accessible ร tous les OFA sous rรฉserve de remplir les 5 conditions suivantes : Dispenser une activitรฉ apprentissage ayant obtenu une certification,Dispenser des formations en apprentissage sur le territoire francilien depuis au moins 1 an en qualitรฉ de CFA, dโOFA ou dโUFA,Prรฉsenter un projet dโinvestissement prรฉvu pour la dispense de formations en apprentissage sur le territoire francilien,รtre propriรฉtaire du bien pour lequel une subvention est sollicitรฉe ou titulaire dโun bail rรฉcemment renouvelรฉ (ou engagement du propriรฉtaire ร renouveler le bail), en propre ou sous la forme de SCI, et assurant la maรฎtrise dโouvrage des travaux dโinvestissement,Prรฉsenter un besoin de financement sur le projet dโinvestissement ne pouvant รชtre pris en charge au titre des fonds propres de la structure et de tiers financeurs1Est-ce que ma structure qui dispense des formations en apprentissage doit avoir une certaine anciennetรฉ pour bรฉnรฉficier de l'aide rรฉgionale?Dispenser des formations en apprentissage sur le territoire francilien depuis au moins 1 an en qualitรฉ de CFA, dโOFA ou dโUFA1Comment dois-je procรฉder pour soumettre ma demande de soutien ร l'investissement pour mon organisme de formation?L'organisme doit dรฉposer sa demande et les piรจces justificatives via le portail mesdemarches.iledefrance.fr1 - Loss:
MultipleNegativesRankingLosswith these parameters:{ "scale": 20.0, "similarity_fct": "cos_sim" }
Training Hyperparameters
Non-Default Hyperparameters
eval_strategy: epochper_device_train_batch_size: 2per_device_eval_batch_size: 2num_train_epochs: 2lr_scheduler_type: cosinewarmup_ratio: 0.1bf16: Truetf32: Trueoptim: adamw_torch_fusedbatch_sampler: no_duplicates
All Hyperparameters
Click to expand
overwrite_output_dir: Falsedo_predict: Falseeval_strategy: epochprediction_loss_only: Trueper_device_train_batch_size: 2per_device_eval_batch_size: 2per_gpu_train_batch_size: Noneper_gpu_eval_batch_size: Nonegradient_accumulation_steps: 1eval_accumulation_steps: Nonetorch_empty_cache_steps: Nonelearning_rate: 5e-05weight_decay: 0.0adam_beta1: 0.9adam_beta2: 0.999adam_epsilon: 1e-08max_grad_norm: 1.0num_train_epochs: 2max_steps: -1lr_scheduler_type: cosinelr_scheduler_kwargs: {}warmup_ratio: 0.1warmup_steps: 0log_level: passivelog_level_replica: warninglog_on_each_node: Truelogging_nan_inf_filter: Truesave_safetensors: Truesave_on_each_node: Falsesave_only_model: Falserestore_callback_states_from_checkpoint: Falseno_cuda: Falseuse_cpu: Falseuse_mps_device: Falseseed: 42data_seed: Nonejit_mode_eval: Falseuse_ipex: Falsebf16: Truefp16: Falsefp16_opt_level: O1half_precision_backend: autobf16_full_eval: Falsefp16_full_eval: Falsetf32: Truelocal_rank: 0ddp_backend: Nonetpu_num_cores: Nonetpu_metrics_debug: Falsedebug: []dataloader_drop_last: Falsedataloader_num_workers: 0dataloader_prefetch_factor: Nonepast_index: -1disable_tqdm: Falseremove_unused_columns: Truelabel_names: Noneload_best_model_at_end: Falseignore_data_skip: Falsefsdp: []fsdp_min_num_params: 0fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}fsdp_transformer_layer_cls_to_wrap: Noneaccelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}deepspeed: Nonelabel_smoothing_factor: 0.0optim: adamw_torch_fusedoptim_args: Noneadafactor: Falsegroup_by_length: Falselength_column_name: lengthddp_find_unused_parameters: Noneddp_bucket_cap_mb: Noneddp_broadcast_buffers: Falsedataloader_pin_memory: Truedataloader_persistent_workers: Falseskip_memory_metrics: Trueuse_legacy_prediction_loop: Falsepush_to_hub: Falseresume_from_checkpoint: Nonehub_model_id: Nonehub_strategy: every_savehub_private_repo: Nonehub_always_push: Falsegradient_checkpointing: Falsegradient_checkpointing_kwargs: Noneinclude_inputs_for_metrics: Falseinclude_for_metrics: []eval_do_concat_batches: Truefp16_backend: autopush_to_hub_model_id: Nonepush_to_hub_organization: Nonemp_parameters:auto_find_batch_size: Falsefull_determinism: Falsetorchdynamo: Noneray_scope: lastddp_timeout: 1800torch_compile: Falsetorch_compile_backend: Nonetorch_compile_mode: Nonedispatch_batches: Nonesplit_batches: Noneinclude_tokens_per_second: Falseinclude_num_input_tokens_seen: Falseneftune_noise_alpha: Noneoptim_target_modules: Nonebatch_eval_metrics: Falseeval_on_start: Falseuse_liger_kernel: Falseeval_use_gather_object: Falseaverage_tokens_across_devices: Falseprompts: Nonebatch_sampler: no_duplicatesmulti_dataset_batch_sampler: proportional
Training Logs
| Epoch | Step | Validation Loss | EmbeddingSimEval_spearman_cosine | BinaryClassifEval_cosine_ap |
|---|---|---|---|---|
| 1.0 | 3 | 0.2267 | nan | 1.0 |
| 2.0 | 6 | 0.2448 | nan | 1.0 |
Framework Versions
- Python: 3.11.9
- Sentence Transformers: 3.4.1
- Transformers: 4.48.3
- PyTorch: 2.3.0
- Accelerate: 1.1.0
- Datasets: 3.3.2
- Tokenizers: 0.21.0
Citation
BibTeX
Sentence Transformers
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}
MultipleNegativesRankingLoss
@misc{henderson2017efficient,
title={Efficient Natural Language Response Suggestion for Smart Reply},
author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
year={2017},
eprint={1705.00652},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
- Downloads last month
- 8
Model tree for romain125/model
Base model
BAAI/bge-base-en-v1.5Evaluation results
- Pearson Cosine on EmbeddingSimEvalself-reportedNaN
- Spearman Cosine on EmbeddingSimEvalself-reportedNaN
- Cosine Accuracy on BinaryClassifEvalself-reported0.800
- Cosine Accuracy Threshold on BinaryClassifEvalself-reported0.653
- Cosine F1 on BinaryClassifEvalself-reported0.889
- Cosine F1 Threshold on BinaryClassifEvalself-reported0.653
- Cosine Precision on BinaryClassifEvalself-reported1.000
- Cosine Recall on BinaryClassifEvalself-reported0.800
- Cosine Ap on BinaryClassifEvalself-reported1.000
- Cosine Mcc on BinaryClassifEvalself-reported0.000