Shakhovak/tiny_sent_transformer_v2

SentenceTransformer based on cointegrated/rubert-tiny2

This is a sentence-transformers model finetuned from cointegrated/rubert-tiny2. It maps sentences & paragraphs to a 312-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: cointegrated/rubert-tiny2
• Maximum Sequence Length: 2048 tokens
• Output Dimensionality: 312 tokens
• Similarity Function: Cosine Similarity

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': 2048, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 312, '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 🤗 Hub
model = SentenceTransformer("sentence_transformers_model_id")
# Run inference
sentences = [
    'Беговая дорожка Hasttings CT100 Главная Беговые дорожки Беговая дорожка Hasttings CT100',
    'Беговая дорожка AMMITY SPACE ATM 5000 Главная Беговые дорожки Бренды',
    'Беговая дорожка ProForm 910 Беговые дорожки ProForm ProForm 910',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 312]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Evaluation

Metrics

Binary Classification

• Dataset: cv
• Evaluated with BinaryClassificationEvaluator

Metric	Value
cosine_accuracy	1.0
cosine_accuracy_threshold	0.7241
cosine_f1	1.0
cosine_f1_threshold	0.7241
cosine_precision	1.0
cosine_recall	1.0
cosine_ap	1.0
dot_accuracy	1.0
dot_accuracy_threshold	0.7241
dot_f1	1.0
dot_f1_threshold	0.7241
dot_precision	1.0
dot_recall	1.0
dot_ap	1.0
manhattan_accuracy	1.0
manhattan_accuracy_threshold	9.0554
manhattan_f1	1.0
manhattan_f1_threshold	9.0554
manhattan_precision	1.0
manhattan_recall	1.0
manhattan_ap	1.0
euclidean_accuracy	1.0
euclidean_accuracy_threshold	0.6519
euclidean_f1	1.0
euclidean_f1_threshold	0.6519
euclidean_precision	1.0
euclidean_recall	1.0
euclidean_ap	1.0
max_accuracy	1.0
max_accuracy_threshold	9.0554
max_f1	1.0
max_f1_threshold	9.0554
max_precision	1.0
max_recall	1.0
max_ap	1.0

Training Details

Training Dataset

Unnamed Dataset

• Size: 13,690 training samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 14 tokens mean: 29.13 tokens max: 66 tokens	min: 13 tokens mean: 29.18 tokens max: 63 tokens	min: 0.0 mean: 0.51 max: 1.0

• Samples:

  sentence1	sentence2	score
  Велотренажер аэродинамический Spirit Fitness AB900+ Air Bike в Москве Кардиотренажеры Велотренажеры Spirit Fitness AB900+ Air Bike	Велотренажер IZHIMIO СL 1500 Главная Велотренажеры Бренды	1.0
  Эллиптический тренажер Sports Art E835 Главная Эллиптические тренажеры Бренды	Степпер Matrix C7XI в Москве Кардиотренажеры Степперы Matrix C7XI	0.0
  Мультистанция Nohrd SlimBeam nan Силовые тренажеры Мультистанции	Эллиптический тренажер Koenigsmann JX-170EF в Москве Кардиотренажеры Эллиптические тренажеры Koenigsmann JX-170EF	0.0

• Loss: ContrastiveLoss with these parameters:

  {
      "distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
      "margin": 0.5,
      "size_average": true
  }
  

Evaluation Dataset

Unnamed Dataset

• Size: 28 evaluation samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 28 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 15 tokens mean: 27.18 tokens max: 45 tokens	min: 16 tokens mean: 28.0 tokens max: 47 tokens	min: 0.0 mean: 0.61 max: 1.0

• Samples:

  sentence1	sentence2	score
  Беговая дорожка Carbon Yukon Беговые дорожки Carbon Carbon Yukon	Беговая дорожка Hasttings LCT80 Беговые дорожки Hasttings Hasttings LCT80	1.0
  Беговая дорожка Беговая дорожка DFC BOSS I T-B1 для реабилитации Беговые дорожки DFC Беговая дорожка DFC BOSS I T-B1 для реабилитации	Беговая дорожка EVO FITNESS Cosmo 5 Главная Беговые дорожки Бренды	1.0
  Грузоблочный тренажер Precor C010ES - жим ногами/икроножные в Москве Силовые тренажеры Грузоблочные Precor C010ES - жим ногами/икроножные	Кроссовер Bronze Gym D-005 Главная Силовые тренажеры Кроссовер Bronze Gym D-005	1.0

• Loss: ContrastiveLoss with these parameters:

  {
      "distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
      "margin": 0.5,
      "size_average": true
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: epoch
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 32
• num_train_epochs: 10
• warmup_ratio: 0.1
• fp16: True

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: epoch
• prediction_loss_only: True
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 32
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 5e-05
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 10
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.1
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• save_safetensors: True
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• no_cuda: False
• use_cpu: False
• use_mps_device: False
• seed: 42
• data_seed: None
• jit_mode_eval: False
• use_ipex: False
• bf16: False
• fp16: True
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: 0
• ddp_backend: None
• tpu_num_cores: None
• tpu_metrics_debug: False
• debug: []
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• past_index: -1
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: []
• fsdp_min_num_params: 0
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• fsdp_transformer_layer_cls_to_wrap: None
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch
• optim_args: None
• adafactor: False
• group_by_length: False
• length_column_name: length
• ddp_find_unused_parameters: None
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• dataloader_pin_memory: True
• dataloader_persistent_workers: False
• skip_memory_metrics: True
• use_legacy_prediction_loop: False
• push_to_hub: False
• resume_from_checkpoint: None
• hub_model_id: None
• hub_strategy: every_save
• hub_private_repo: False
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• eval_do_concat_batches: True
• fp16_backend: auto
• push_to_hub_model_id: None
• push_to_hub_organization: None
• mp_parameters:
• auto_find_batch_size: False
• full_determinism: False
• torchdynamo: None
• ray_scope: last
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• dispatch_batches: None
• split_batches: None
• include_tokens_per_second: False
• include_num_input_tokens_seen: False
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• eval_on_start: False
• eval_use_gather_object: False
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	loss	cv_max_ap
0	0	-	-	0.7655
1.0	428	-	0.0056	1.0
1.1682	500	0.0078	-	-
2.0	856	-	0.0015	1.0
2.3364	1000	0.0019	-	-
3.0	1284	-	0.0011	1.0
3.5047	1500	0.0013	-	-
4.0	1712	-	0.0007	1.0
4.6729	2000	0.001	-	-
5.0	2140	-	0.0004	1.0
5.8411	2500	0.0008	-	-
6.0	2568	-	0.0002	1.0
7.0	2996	-	0.0002	1.0
7.0093	3000	0.0007	-	-
8.0	3424	-	0.0001	1.0
8.1776	3500	0.0006	-	-
9.0	3852	-	0.0001	1.0
9.3458	4000	0.0005	-	-
10.0	4280	-	0.0001	1.0

Framework Versions

• Python: 3.11.8
• Sentence Transformers: 3.1.0
• Transformers: 4.44.2
• PyTorch: 2.4.1+cu118
• Accelerate: 0.34.2
• Datasets: 3.0.0
• Tokenizers: 0.19.1

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",
}

ContrastiveLoss

@inproceedings{hadsell2006dimensionality,
    author={Hadsell, R. and Chopra, S. and LeCun, Y.},
    booktitle={2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06)},
    title={Dimensionality Reduction by Learning an Invariant Mapping},
    year={2006},
    volume={2},
    number={},
    pages={1735-1742},
    doi={10.1109/CVPR.2006.100}
}