README.md

metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:78879
  - loss:CosineSimilarityLoss
base_model: intfloat/multilingual-e5-base
widget:
  - source_sentence: Somatotropin Ab
    sentences:
      - Desethylamiodarone
      - Glucose^7H post XXX challenge
      - Somatotropin Ab
  - source_sentence: Erythrocytes.fetal/1000 erythrocytes
    sentences:
      - levoFLOXacin
      - Pathologist interpretation
      - Pepsinogen I
  - source_sentence: Aggregazione piastrinica.arachidonato indotta
    sentences:
      - Epidermal growth factor
      - Bilirubin.glucuronidated/Bilirubin.total
      - Platelet aggregation.arachidonate induced
  - source_sentence: Parathormoon.intact^5 min na uitsnijding  in serum of plasma
    sentences:
      - Fatty acids.very long chain
      - Estradiol^4th specimen post XXX challenge
      - Parathyrin.intact^5M post excision
  - source_sentence: Karboksühemoglobiin/hemoglobiin.üld
    sentences:
      - Ammonia
      - Carboxyhemoglobin/Hemoglobin.total
      - Procainamide+N-acetylprocainamide
pipeline_tag: sentence-similarity
library_name: sentence-transformers

SentenceTransformer based on intfloat/multilingual-e5-base

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-base. 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: intfloat/multilingual-e5-base
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 dimensions
• 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': 512, 'do_lower_case': False}) with Transformer model: XLMRobertaModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, '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("iddqd21/fine-tuned-e5-semantic-similarity")
# Run inference
sentences = [
    'Karboksühemoglobiin/hemoglobiin.üld',
    'Carboxyhemoglobin/Hemoglobin.total',
    'Procainamide+N-acetylprocainamide',
]
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]

Training Details

Training Dataset

Unnamed Dataset

• Size: 78,879 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	float
  details	min: 3 tokens mean: 11.64 tokens max: 36 tokens	min: 3 tokens mean: 10.26 tokens max: 32 tokens	min: 0.0 mean: 0.59 max: 1.0

• Samples:

  sentence_0	sentence_1	label
  Rakud.CD3+HLA-DR+/100 raku kohta	Cells.CD3+HLA-DR+/100 cells	1.0
  Zellen.FMC7/100 Zellen	Cells.FMC7/100 cells	1.0
  Apolipoprotéine AI/apolipoprotéine B	Apolipoprotein A-I/Apolipoprotein B	1.0

• Loss: CosineSimilarityLoss with these parameters:

  {
      "loss_fct": "torch.nn.modules.loss.MSELoss"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• num_train_epochs: 10
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: no
• prediction_loss_only: True
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• 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: 10
• 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
• 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
• 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
• use_liger_kernel: False
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: round_robin

Training Logs

Epoch	Step	Training Loss
0.1014	500	0.0633
0.2028	1000	0.0332
0.3043	1500	0.0296
0.4057	2000	0.0266
0.5071	2500	0.024
0.6085	3000	0.0239
0.7099	3500	0.0216
0.8114	4000	0.0205
0.9128	4500	0.0187
1.0142	5000	0.0185
1.1156	5500	0.0149
1.2170	6000	0.015
1.3185	6500	0.0142
1.4199	7000	0.0152
1.5213	7500	0.0138
1.6227	8000	0.0131
1.7241	8500	0.014
1.8256	9000	0.0133
1.9270	9500	0.0125
2.0284	10000	0.0128
2.1298	10500	0.0093
2.2312	11000	0.0091
2.3327	11500	0.0097
2.4341	12000	0.0096
2.5355	12500	0.0097
2.6369	13000	0.0093
2.7383	13500	0.0099
2.8398	14000	0.0104
2.9412	14500	0.009
3.0426	15000	0.0084
3.1440	15500	0.0065
3.2454	16000	0.0062
3.3469	16500	0.0062
3.4483	17000	0.0068
3.5497	17500	0.0076
3.6511	18000	0.0078
3.7525	18500	0.0068
3.8540	19000	0.008
3.9554	19500	0.0076
4.0568	20000	0.0057
4.1582	20500	0.0054
4.2596	21000	0.0052
4.3611	21500	0.0052
4.4625	22000	0.0056
4.5639	22500	0.0055
4.6653	23000	0.0057
4.7667	23500	0.006
4.8682	24000	0.0054
4.9696	24500	0.0052
5.0710	25000	0.0045
5.1724	25500	0.0039
5.2738	26000	0.0043
5.3753	26500	0.004
5.4767	27000	0.0044
5.5781	27500	0.0045
5.6795	28000	0.0039
5.7809	28500	0.0043
5.8824	29000	0.0047
5.9838	29500	0.0049
6.0852	30000	0.003
6.1866	30500	0.0034
6.2880	31000	0.003
6.3895	31500	0.0031
6.4909	32000	0.0033
6.5923	32500	0.0035
6.6937	33000	0.0037
6.7951	33500	0.0039
6.8966	34000	0.004
6.9980	34500	0.003
7.0994	35000	0.0024
7.2008	35500	0.0026
7.3022	36000	0.0029
7.4037	36500	0.0029
7.5051	37000	0.0025
7.6065	37500	0.0026
7.7079	38000	0.0032
7.8093	38500	0.0032
7.9108	39000	0.0029
8.0122	39500	0.0028
8.1136	40000	0.0024
8.2150	40500	0.0021
8.3164	41000	0.0022
8.4178	41500	0.0022
8.5193	42000	0.0024
8.6207	42500	0.0025
8.7221	43000	0.0023
8.8235	43500	0.0021
8.9249	44000	0.0026
9.0264	44500	0.0025
9.1278	45000	0.0021
9.2292	45500	0.0017
9.3306	46000	0.0022
9.4320	46500	0.002
9.5335	47000	0.0021
9.6349	47500	0.0019
9.7363	48000	0.0021
9.8377	48500	0.002
9.9391	49000	0.0021

Framework Versions

• Python: 3.9.20
• Sentence Transformers: 3.3.1
• Transformers: 4.47.1
• PyTorch: 2.5.1+rocm6.2
• Accelerate: 1.2.1
• Datasets: 3.2.0
• 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",
}