README.md

metadata

tags:
  - sentence-transformers
  - cross-encoder
  - reranker
  - generated_from_trainer
  - dataset_size:87398
  - loss:CrossEntropyLoss
base_model: deepvk/USER-bge-m3
pipeline_tag: text-classification
library_name: sentence-transformers
metrics:
  - f1_macro
  - f1_micro
  - f1_weighted
model-index:
  - name: CrossEncoder based on deepvk/USER-bge-m3
    results:
      - task:
          type: cross-encoder-softmax-accuracy
          name: Cross Encoder Softmax Accuracy
        dataset:
          name: softmax accuracy eval
          type: softmax_accuracy_eval
        metrics:
          - type: f1_macro
            value: 0.9726781946411447
            name: F1 Macro
          - type: f1_micro
            value: 0.9753250742295485
            name: F1 Micro
          - type: f1_weighted
            value: 0.9752858934461676
            name: F1 Weighted

CrossEncoder based on deepvk/USER-bge-m3

This is a Cross Encoder model finetuned from deepvk/USER-bge-m3 using the sentence-transformers library. It computes scores for pairs of texts, which can be used for text pair classification.

Model Details

Model Description

• Model Type: Cross Encoder
• Base model: deepvk/USER-bge-m3
• Maximum Sequence Length: 8192 tokens
• Number of Output Labels: 2 labels

Model Sources

• Documentation: Sentence Transformers Documentation
• Documentation: Cross Encoder Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Cross Encoders on Hugging Face

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 CrossEncoder

# Download from the 🤗 Hub
model = CrossEncoder("Chimalpopoka/CrossEncoderRanker")
# Get scores for pairs of texts
pairs = [
    ['Свободный кортизол, суточная\xa0моча (Free Сortisol, Free Hydrocortisone, 24-Hour urine)', 'Кортизол в разовой порции мочи'],
    ['Определение антител класса G (IgG) к RBD домену S белка вируса SARS-CoV-2 (COVID-19), количественное исследование', 'Антитела к миокарду, IgG'],
    ['Прием (осмотр, консультация) врача-терапевта, первичный', 'Консультация врача, в клинике, терапевт'],
    ['Вакцинация против гепатита В для взрослых', 'Вакцинация против гепатита А. Вакцина: Альгавак М (Россия)'],
    ['АТ к миокарду', 'Антитела к миокарду, IgG'],
]
scores = model.predict(pairs)
print(scores.shape)
# (5, 2)

Evaluation

Metrics

Cross Encoder Softmax Accuracy

• Dataset: softmax_accuracy_eval
• Evaluated with CESoftmaxAccuracyEvaluator

Metric	Value
f1_macro	0.9727
f1_micro	0.9753
f1_weighted	0.9753

Training Details

Training Dataset

Unnamed Dataset

• Size: 87,398 training samples
• Columns: sentence_0, sentence_1, and label
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1	label
  type	string	string	int
  details	min: 4 characters mean: 68.59 characters max: 747 characters	min: 3 characters mean: 63.11 characters max: 281 characters	0: ~34.30% 1: ~65.70%

• Samples:

  sentence_0	sentence_1	label
  Свободный кортизол, суточная моча (Free Сortisol, Free Hydrocortisone, 24-Hour urine)	Кортизол в разовой порции мочи	1
  Определение антител класса G (IgG) к RBD домену S белка вируса SARS-CoV-2 (COVID-19), количественное исследование	Антитела к миокарду, IgG	0
  Прием (осмотр, консультация) врача-терапевта, первичный	Консультация врача, в клинике, терапевт	1

• Loss: CrossEntropyLoss

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• num_train_epochs: 1

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 8
• per_device_eval_batch_size: 8
• 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
• num_train_epochs: 1
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.0
• 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: False
• 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: None
• hub_always_push: False
• hub_revision: None
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• include_for_metrics: []
• 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
• 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
• use_liger_kernel: False
• liger_kernel_config: None
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	softmax_accuracy_eval_f1_macro
0.0458	500	0.5651	-
0.0915	1000	0.2182	-
0.1373	1500	0.2239	-
0.1831	2000	0.2015	0.9616
0.2288	2500	0.1617	-
0.2746	3000	0.1942	-
0.3204	3500	0.1888	-
0.3661	4000	0.1772	0.9629
0.4119	4500	0.1635	-
0.4577	5000	0.1596	-
0.5034	5500	0.1709	-
0.5492	6000	0.1566	0.9640
0.5950	6500	0.1278	-
0.6407	7000	0.1276	-
0.6865	7500	0.1339	-
0.7323	8000	0.1422	0.9715
0.7780	8500	0.148	-
0.8238	9000	0.1271	-
0.8696	9500	0.125	-
0.9153	10000	0.1103	0.9727

Framework Versions

• Python: 3.12.3
• Sentence Transformers: 5.1.0
• Transformers: 4.53.2
• PyTorch: 2.7.1+cu126
• Accelerate: 1.10.1
• Datasets: 4.0.0
• Tokenizers: 0.21.2

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",
}