Fallovski/CrossencoderPharma

CrossEncoder based on doctolib-lab/doctobert-fr-base

This is a Cross Encoder model finetuned from doctolib-lab/doctobert-fr-base using the sentence-transformers library. It computes scores for pairs of texts, which can be used for text reranking and semantic search.

Model Details

Model Description

• Model Type: Cross Encoder
• Base model: doctolib-lab/doctobert-fr-base
• Maximum Sequence Length: 64 tokens
• Number of Output Labels: 1 label
• Supported Modality: Text

Model Sources

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

Full Model Architecture

CrossEncoder(
  (0): Transformer({'transformer_task': 'sequence-classification', 'modality_config': {'text': {'method': 'forward', 'method_output_name': 'logits'}}, 'module_output_name': 'scores', 'architecture': 'RobertaForSequenceClassification'})
)

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("Fallovski/CrossencoderPharma")
# Get scores for pairs of inputs
pairs = [
    ['KLIPAL CODEINE 300/25 MG CP /16', 'ANTHELIOS 30+ ULT YX SENS 50ML'],
    ['BIOTIC PLUS 100MG ENF 60ML', 'BIOTIC PLUS 500MG SACH B/14'],
    ['FALCIART 80/480 MG COMP B/6', 'FALCIART CPR 80/480MG BT6'],
    ['KLACIN CPR 625MG BT15', 'Klacin Cpr 625Mg Bt15'],
    ['INDOCOLLYRE 0.1% ETO F/5ML B/1', 'Indocollyre 0.1% Eto F/5Ml B/1'],
]
scores = model.predict(pairs)
print(scores)
# [9.7995e-04 1.1121e-03 9.9875e-01 9.9884e-01 9.9885e-01]

# Or rank different texts based on similarity to a single text
ranks = model.rank(
    'KLIPAL CODEINE 300/25 MG CP /16',
    [
        'ANTHELIOS 30+ ULT YX SENS 50ML',
        'BIOTIC PLUS 500MG SACH B/14',
        'FALCIART CPR 80/480MG BT6',
        'Klacin Cpr 625Mg Bt15',
        'Indocollyre 0.1% Eto F/5Ml B/1',
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

Evaluation

Metrics

Cross Encoder Classification

• Dataset: validation
• Evaluated with CrossEncoderClassificationEvaluator

Metric	Value
accuracy	0.9871
accuracy_threshold	0.5495
f1	0.9869
f1_threshold	0.5495
precision	0.9849
recall	0.9889
average_precision	0.9981

Training Details

Training Dataset

Unnamed Dataset

• Size: 11,426 training samples
• Columns: sentence1, sentence2, and label
• Approximate statistics based on the first 100 samples:

  	sentence1	sentence2	label
  type	string	string	float
  modality	text	text
  details	min: 6 tokens mean: 16.49 tokens max: 27 tokens	min: 5 tokens mean: 16.99 tokens max: 28 tokens	min: 0.0 mean: 0.51 max: 1.0

• Samples:

  sentence1	sentence2	label
  SÉROLOGIE ARBOVIRUS	SÉROLOGIE WRIGHT	0.0
  SOND FOL H 15ML AA14 16CH /1	SOND FOL H 15ML AA14 16CH 1	1.0
  FARLINE CAPRI D +1.5 MARON BLANC HOM	FARLINE MONZA CAREY +3.0 MARON/NOIR	0.0

• Loss: BinaryCrossEntropyLoss with these parameters:

  {
      "activation_fn": "torch.nn.modules.linear.Identity",
      "pos_weight": null
  }
  

Evaluation Dataset

Unnamed Dataset

• Size: 2,016 evaluation samples
• Columns: sentence1, sentence2, and label
• Approximate statistics based on the first 100 samples:

  	sentence1	sentence2	label
  type	string	string	float
  modality	text	text
  details	min: 5 tokens mean: 16.54 tokens max: 35 tokens	min: 5 tokens mean: 16.98 tokens max: 33 tokens	min: 0.0 mean: 0.51 max: 1.0

• Samples:

  sentence1	sentence2	label
  KLIPAL CODEINE 300/25 MG CP /16	ANTHELIOS 30+ ULT YX SENS 50ML	0.0
  BIOTIC PLUS 100MG ENF 60ML	BIOTIC PLUS 500MG SACH B/14	0.0
  FALCIART 80/480 MG COMP B/6	FALCIART CPR 80/480MG BT6	1.0

• Loss: BinaryCrossEntropyLoss with these parameters:

  {
      "activation_fn": "torch.nn.modules.linear.Identity",
      "pos_weight": null
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size: 32
• per_device_eval_batch_size: 32
• learning_rate: 2e-05
• num_train_epochs: 6
• warmup_ratio: 0.1
• load_best_model_at_end: True

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• 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: 2e-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: 6
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: None
• 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
• 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: True
• 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}
• parallelism_config: None
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch_fused
• optim_args: None
• adafactor: False
• group_by_length: False
• length_column_name: length
• project: huggingface
• trackio_space_id: trackio
• 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: no
• 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: True
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss	validation_average_precision
0.1397	50	0.6869	-	-
0.2793	100	0.3572	-	-
0.4190	150	0.1483	-	-
0.5587	200	0.13	-	-
0.6983	250	0.1394	-	-
0.8380	300	0.106	-	-
0.9777	350	0.0999	-	-
1.0	358	-	0.1673	0.9963
1.1173	400	0.084	-	-
1.2570	450	0.0731	-	-
1.3966	500	0.0716	-	-
1.5363	550	0.0699	-	-
1.6760	600	0.0647	-	-
1.8156	650	0.0691	-	-
1.9553	700	0.0942	-	-
2.0	716	-	0.0571	0.9981
2.0950	750	0.0399	-	-
2.2346	800	0.0503	-	-
2.3743	850	0.0219	-	-
2.5140	900	0.0456	-	-
2.6536	950	0.0536	-	-
2.7933	1000	0.0349	-	-
2.9330	1050	0.052	-	-
3.0	1074	-	0.0602	0.9983
3.0726	1100	0.0398	-	-
3.2123	1150	0.0128	-	-
3.3520	1200	0.022	-	-
3.4916	1250	0.0122	-	-
3.6313	1300	0.0265	-	-
3.7709	1350	0.04	-	-
3.9106	1400	0.0242	-	-
4.0	1432	-	0.0684	0.9987
4.0503	1450	0.0186	-	-
4.1899	1500	0.0185	-	-
4.3296	1550	0.0129	-	-
4.4693	1600	0.0113	-	-
4.6089	1650	0.0103	-	-
4.7486	1700	0.003	-	-
4.8883	1750	0.0207	-	-
5.0	1790	-	0.0744	0.9987
5.0279	1800	0.0181	-	-
5.1676	1850	0.0158	-	-
5.3073	1900	0.0156	-	-
5.4469	1950	0.0123	-	-
5.5866	2000	0.0094	-	-
5.7263	2050	0.0159	-	-
5.8659	2100	0.0033	-	-
6.0	2148	-	0.0649	0.9988
-1	-1	-	-	0.9981

• The bold row denotes the saved checkpoint.

Training Time

• Training: 10.8 minutes
• Evaluation: 1.1 minutes
• Total: 11.9 minutes

Framework Versions

• Python: 3.13.11
• Sentence Transformers: 5.6.1
• Transformers: 4.57.6
• PyTorch: 2.13.0+cu130
• Accelerate: 1.14.0
• Datasets: 5.0.1
• Tokenizers: 0.22.2

Additional Resources

• Training and Finetuning Reranker Models with Sentence Transformers: the end-to-end guide for training or finetuning Cross Encoder (reranker) models.
• Multimodal Embedding & Reranker Models with Sentence Transformers: use text, image, audio, and video reranker models through the same API.
• Training and Finetuning Multimodal Embedding & Reranker Models with Sentence Transformers: training multimodal Cross Encoders.

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",
}