Hugging Face's logo

hilabs
/
automapper_molina

Sentence Similarity
sentence-transformers
Safetensors
English
bert
feature-extraction
dense
Generated from Trainer
dataset_size:91097
loss:MultipleNegativesRankingLoss
Eval Results (legacy)
text-embeddings-inference
Model card Files
xet
Community

txt_std_ra_automapper_molina

This is a sentence-transformers model finetuned from NeuML/pubmedbert-base-embeddings. 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: NeuML/pubmedbert-base-embeddings
  • Maximum Sequence Length: 512 tokens
  • Output Dimensionality: 768 dimensions
  • Similarity Function: Cosine Similarity
  • Language: en
  • License: apache-2.0

Model Sources

Full Model Architecture 

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False, 'architecture': 'BertModel'})
  (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})
)

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 = [
    '<DOMAIN>ROSTER<TASK>COL_MAP<TEXT> affiliate address<VALUE>200 E ARIZONA||3480 E GUASTI RD',
    '<DOMAIN>ROSTER<TASK>COL_MAP<TEXT>Hospital Addr Line 1',
    '<DOMAIN>ROSTER<TASK>COL_MAP<TEXT>Hospital State',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.9630, 0.0580],
#         [0.9630, 1.0000, 0.0549],
#         [0.0580, 0.0549, 1.0000]])

Evaluation

Metrics

Triplet

Metric Value
cosine_accuracy 0.9426

Training Details

Training Dataset

Unnamed Dataset

  • Size: 91,097 training samples
  • Columns: anchor, positive, and negative
  • Approximate statistics based on the first 1000 samples:
    anchor positive negative
    type string string string
    details
    • min: 21 tokens
    • mean: 45.98 tokens
    • max: 185 tokens
    • min: 17 tokens
    • mean: 19.65 tokens
    • max: 30 tokens
    • min: 17 tokens
    • mean: 19.86 tokens
    • max: 30 tokens
  • Samples:
    anchor positive negative
    ROSTERCOL_MAP ROSTERCOL_MAP ROSTERCOL_MAP
    ROSTERCOL_MAP ROSTERCOL_MAP ROSTERCOL_MAP
    ROSTERCOL_MAP ROSTERCOL_MAP ROSTERCOL_MAP
  • Loss: MultipleNegativesRankingLoss with these parameters:
    {
    "scale": 20.0,
    "similarity_fct": "cos_sim",
    "gather_across_devices": false
}

Evaluation Dataset

Unnamed Dataset

  • Size: 9,709 evaluation samples
  • Columns: anchor, positive, and negative
  • Approximate statistics based on the first 1000 samples:
    anchor positive negative
    type string string string
    details
    • min: 21 tokens
    • mean: 46.15 tokens
    • max: 253 tokens
    • min: 17 tokens
    • mean: 19.7 tokens
    • max: 30 tokens
    • min: 17 tokens
    • mean: 19.94 tokens
    • max: 30 tokens
  • Samples:
    anchor positive negative
    ROSTERCOL_MAP ROSTERCOL_MAP ROSTERCOL_MAP
    ROSTERCOL_MAP ROSTERCOL_MAP ROSTERCOL_MAP
    ROSTERCOL_MAP ROSTERCOL_MAP ROSTERCOL_MAP
  • Loss: MultipleNegativesRankingLoss with these parameters:
    {
    "scale": 20.0,
    "similarity_fct": "cos_sim",
    "gather_across_devices": false
}

Training Hyperparameters

Non-Default Hyperparameters

  • eval_strategy: steps
  • per_device_train_batch_size: 16
  • per_device_eval_batch_size: 16
  • num_train_epochs: 2
  • warmup_ratio: 0.1
  • fp16: True
  • load_best_model_at_end: True
  • batch_sampler: no_duplicates

All Hyperparameters

Click to expand
  • overwrite_output_dir: False
  • do_predict: False
  • eval_strategy: steps
  • 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: 1.0
  • num_train_epochs: 2
  • 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: 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}
  • deepspeed: None
  • label_smoothing_factor: 0.0
  • optim: adamw_torch_fused
  • 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: no_duplicates
  • multi_dataset_batch_sampler: proportional
  • router_mapping: {}
  • learning_rate_mapping: {}

Training Logs

Click to expand
Epoch Step Training Loss Validation Loss txt_std_ra_automapper_molina_cosine_accuracy
-1 -1 - - 0.9426
0.0088 50 1.7127 - -
0.0176 100 0.8476 - -
0.0263 150 0.5572 - -
0.0351 200 0.4306 0.2925 -
0.0439 250 0.3464 - -
0.0527 300 0.2764 - -
0.0615 350 0.2381 - -
0.0702 400 0.1622 0.1312 -
0.0790 450 0.1617 - -
0.0878 500 0.1492 - -
0.0966 550 0.1429 - -
0.1054 600 0.114 0.0835 -
0.1142 650 0.1203 - -
0.1229 700 0.0901 - -
0.1317 750 0.1014 - -
0.1405 800 0.0796 0.0614 -
0.1493 850 0.0631 - -
0.1581 900 0.0989 - -
0.1668 950 0.0627 - -
0.1756 1000 0.0809 0.0670 -
0.1844 1050 0.0638 - -
0.1932 1100 0.0664 - -
0.2020 1150 0.0419 - -
0.2107 1200 0.0569 0.0506 -
0.2195 1250 0.0842 - -
0.2283 1300 0.0557 - -
0.2371 1350 0.0653 - -
0.2459 1400 0.065 0.0438 -
0.2547 1450 0.0459 - -
0.2634 1500 0.0644 - -
0.2722 1550 0.0494 - -
0.2810 1600 0.0532 0.0296 -
0.2898 1650 0.0792 - -
0.2986 1700 0.0592 - -
0.3073 1750 0.0503 - -
0.3161 1800 0.0353 0.0375 -
0.3249 1850 0.0556 - -
0.3337 1900 0.0545 - -
0.3425 1950 0.035 - -
0.3512 2000 0.0286 0.0219 -
0.3600 2050 0.0305 - -
0.3688 2100 0.014 - -
0.3776 2150 0.0307 - -
0.3864 2200 0.0374 0.0242 -
0.3952 2250 0.039 - -
0.4039 2300 0.0218 - -
0.4127 2350 0.0416 - -
0.4215 2400 0.038 0.0186 -
0.4303 2450 0.0282 - -
0.4391 2500 0.0171 - -
0.4478 2550 0.0282 - -
0.4566 2600 0.0227 0.0185 -
0.4654 2650 0.02 - -
0.4742 2700 0.0215 - -
0.4830 2750 0.0328 - -
0.4917 2800 0.0118 0.0165 -
0.5005 2850 0.0278 - -
0.5093 2900 0.0072 - -
0.5181 2950 0.0252 - -
0.5269 3000 0.0162 0.0151 -
0.5357 3050 0.0241 - -
0.5444 3100 0.0042 - -
0.5532 3150 0.0157 - -
0.5620 3200 0.0256 0.0141 -
0.5708 3250 0.0106 - -
0.5796 3300 0.0138 - -
0.5883 3350 0.0292 - -
0.5971 3400 0.0133 0.0164 -
0.6059 3450 0.0105 - -
0.6147 3500 0.0148 - -
0.6235 3550 0.0101 - -
0.6322 3600 0.0101 0.0136 -
0.6410 3650 0.0271 - -
0.6498 3700 0.028 - -
0.6586 3750 0.0057 - -
0.6674 3800 0.0273 0.0101 -
0.6762 3850 0.0201 - -
0.6849 3900 0.0164 - -
0.6937 3950 0.0425 - -
0.7025 4000 0.0168 0.0112 -
0.7113 4050 0.0174 - -
0.7201 4100 0.0153 - -
0.7288 4150 0.0166 - -
0.7376 4200 0.0252 0.0078 -
0.7464 4250 0.0098 - -
0.7552 4300 0.0145 - -
0.7640 4350 0.0141 - -
0.7727 4400 0.0119 0.0088 -
0.7815 4450 0.0108 - -
0.7903 4500 0.0146 - -
0.7991 4550 0.0104 - -
0.8079 4600 0.0068 0.0116 -
0.8166 4650 0.0233 - -
0.8254 4700 0.0028 - -
0.8342 4750 0.0255 - -
0.8430 4800 0.009 0.0127 -
0.8518 4850 0.0293 - -
0.8606 4900 0.0045 - -
0.8693 4950 0.0048 - -
0.8781 5000 0.0178 0.0132 -
0.8869 5050 0.0059 - -
0.8957 5100 0.0221 - -
0.9045 5150 0.0082 - -
0.9132 5200 0.0111 0.0097 -
0.9220 5250 0.0021 - -
0.9308 5300 0.0034 - -
0.9396 5350 0.0449 - -
0.9484 5400 0.0128 0.0066 -
0.9571 5450 0.0095 - -
0.9659 5500 0.009 - -
0.9747 5550 0.0169 - -
0.9835 5600 0.0115 0.0060 -
0.9923 5650 0.0204 - -
1.0011 5700 0.0116 - -
1.0098 5750 0.0049 - -
1.0186 5800 0.0064 0.0096 -
1.0274 5850 0.0061 - -
1.0362 5900 0.0011 - -
1.0450 5950 0.018 - -
1.0537 6000 0.0231 0.0056 -
1.0625 6050 0.0081 - -
1.0713 6100 0.0021 - -
1.0801 6150 0.006 - -
1.0889 6200 0.0116 0.0078 -
1.0976 6250 0.0074 - -
1.1064 6300 0.0082 - -
1.1152 6350 0.0011 - -
1.1240 6400 0.0051 0.0101 -
1.1328 6450 0.007 - -
1.1416 6500 0.0015 - -
1.1503 6550 0.0037 - -
1.1591 6600 0.0027 0.0073 -
1.1679 6650 0.0005 - -
1.1767 6700 0.0239 - -
1.1855 6750 0.0136 - -
1.1942 6800 0.0251 0.0070 -
1.2030 6850 0.0004 - -
1.2118 6900 0.0065 - -
1.2206 6950 0.0109 - -
1.2294 7000 0.0009 0.0043 -
1.2381 7050 0.0086 - -
1.2469 7100 0.003 - -
1.2557 7150 0.0044 - -
1.2645 7200 0.0118 0.0058 -
1.2733 7250 0.0093 - -
1.2821 7300 0.0023 - -
1.2908 7350 0.002 - -
1.2996 7400 0.007 0.0061 -
1.3084 7450 0.0162 - -
1.3172 7500 0.0011 - -
1.3260 7550 0.007 - -
1.3347 7600 0.0014 0.0048 -
1.3435 7650 0.0033 - -
1.3523 7700 0.0007 - -
1.3611 7750 0.0017 - -
1.3699 7800 0.0078 0.0049 -
1.3786 7850 0.0049 - -
1.3874 7900 0.003 - -
1.3962 7950 0.0028 - -
1.4050 8000 0.0038 0.0033 -
1.4138 8050 0.0158 - -
1.4226 8100 0.0008 - -
1.4313 8150 0.0007 - -
1.4401 8200 0.0038 0.0024 -
1.4489 8250 0.0177 - -
1.4577 8300 0.0044 - -
1.4665 8350 0.0064 - -
1.4752 8400 0.0005 0.0049 -
1.4840 8450 0.0146 - -
1.4928 8500 0.001 - -
1.5016 8550 0.0014 - -
1.5104 8600 0.0041 0.0038 -
1.5191 8650 0.0072 - -
1.5279 8700 0.0014 - -
1.5367 8750 0.0135 - -
1.5455 8800 0.0148 0.0039 -
1.5543 8850 0.0017 - -
1.5630 8900 0.007 - -
1.5718 8950 0.012 - -
1.5806 9000 0.0004 0.0024 -
1.5894 9050 0.0026 - -
1.5982 9100 0.0109 - -
1.6070 9150 0.0009 - -
1.6157 9200 0.0054 0.0022 -
1.6245 9250 0.0032 - -
1.6333 9300 0.0135 - -
1.6421 9350 0.0131 - -
1.6509 9400 0.0049 0.0021 -
1.6596 9450 0.0003 - -
1.6684 9500 0.0027 - -
1.6772 9550 0.0008 - -
1.686 9600 0.0124 0.002 -
1.6948 9650 0.0026 - -
1.7035 9700 0.004 - -
1.7123 9750 0.0008 - -
1.7211 9800 0.0058 0.0028 -
1.7299 9850 0.0133 - -
1.7387 9900 0.0005 - -
1.7475 9950 0.0007 - -
1.7562 10000 0.0009 0.0040 -
1.7650 10050 0.0018 - -
1.7738 10100 0.0002 - -
1.7826 10150 0.002 - -
1.7914 10200 0.0021 0.0042 -
1.8001 10250 0.002 - -
1.8089 10300 0.0003 - -
1.8177 10350 0.001 - -
1.8265 10400 0.0092 0.0045 -
1.8353 10450 0.0032 - -
1.8440 10500 0.0002 - -
1.8528 10550 0.0026 - -
1.8616 10600 0.0003 0.0048 -
1.8704 10650 0.001 - -
1.8792 10700 0.0126 - -
1.8880 10750 0.0172 - -
1.8967 10800 0.0002 0.0034 -
1.9055 10850 0.0038 - -
1.9143 10900 0.0005 - -
1.9231 10950 0.0001 - -
1.9319 11000 0.0162 0.0033 -
1.9406 11050 0.0037 - -
1.9494 11100 0.0003 - -
1.9582 11150 0.0021 - -
1.9670 11200 0.0098 0.0032 -
1.9758 11250 0.0003 - -
1.9845 11300 0.0017 - -
1.9933 11350 0.0048 - -
  • The bold row denotes the saved checkpoint.

Framework Versions

  • Python: 3.10.17
  • Sentence Transformers: 5.1.0
  • Transformers: 4.55.0
  • PyTorch: 2.8.0+cu128
  • Accelerate: 1.7.0
  • Datasets: 3.5.0
  • Tokenizers: 0.21.4

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
12
Safetensors
Model size
0.1B params
Tensor type
F32
·
Inference Providers NEW
Sentence Similarity
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for hilabs/automapper_molina

Base model

microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext
Finetuned
NeuML/pubmedbert-base-embeddings
Finetuned
(13)
this model

Papers for hilabs/automapper_molina

Evaluation results