| --- |
| library_name: transformers |
| language: |
| - en |
| license: apache-2.0 |
| base_model: microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext |
| tags: |
| - generated_from_trainer |
| metrics: |
| - precision |
| - recall |
| - accuracy |
| model-index: |
| - name: CeLLaTe-ner-2class-pubmedbert-baseline |
| results: [] |
| --- |
| |
| <!-- This model card has been generated automatically according to the information the Trainer had access to. You |
| should probably proofread and complete it, then remove this comment. --> |
|
|
| # CeLLaTe-ner-2class-pubmedbert-baseline |
|
|
| This model is a fine-tuned version of [microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext](https://huggingface.co/microsoft/BiomedNLP-BiomedBERT-base-uncased-abstract-fulltext) on the OTAR3088/CeLLaTe-ner-2class-iob_final dataset. |
| It achieves the following results on the evaluation set: |
| - Loss: 0.1003 |
| - Precision: 0.7763 |
| - Recall: 0.7431 |
| - Micro F1: 0.7594 |
| - Weighted F1: 0.7596 |
| - Macro F1: 0.7690 |
| - Accuracy: 0.9838 |
| |
| ## Model description |
| |
| More information needed |
| |
| ## Intended uses & limitations |
| |
| More information needed |
| |
| ## Training and evaluation data |
| |
| More information needed |
| |
| ## Training procedure |
| |
| ### Training hyperparameters |
| |
| The following hyperparameters were used during training: |
| - learning_rate: 2e-05 |
| - train_batch_size: 16 |
| - eval_batch_size: 16 |
| - seed: 3407 |
| - gradient_accumulation_steps: 2 |
| - total_train_batch_size: 32 |
| - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
| - lr_scheduler_type: linear |
| - lr_scheduler_warmup_ratio: 0.01 |
| - num_epochs: 20 |
| - mixed_precision_training: Native AMP |
| |
| ### Training results |
| |
| | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | Micro F1 | Weighted F1 | Macro F1 | Accuracy | |
| |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:--------:|:-----------:|:--------:|:--------:| |
| | 0.2498 | 1.0 | 263 | 0.0637 | 0.6194 | 0.7517 | 0.6792 | 0.6827 | 0.7051 | 0.9791 | |
| | 0.0412 | 2.0 | 526 | 0.0568 | 0.7413 | 0.7114 | 0.7260 | 0.7269 | 0.7388 | 0.9823 | |
| | 0.0268 | 3.0 | 789 | 0.0633 | 0.7183 | 0.6785 | 0.6978 | 0.6978 | 0.6973 | 0.9809 | |
| | 0.0199 | 4.0 | 1052 | 0.0655 | 0.7478 | 0.7200 | 0.7336 | 0.7345 | 0.7464 | 0.9826 | |
| | 0.0148 | 5.0 | 1315 | 0.0713 | 0.7721 | 0.7029 | 0.7359 | 0.7358 | 0.7406 | 0.9828 | |
| | 0.0119 | 6.0 | 1578 | 0.0724 | 0.7591 | 0.7364 | 0.7476 | 0.7480 | 0.7569 | 0.9833 | |
| | 0.0091 | 7.0 | 1841 | 0.0867 | 0.7860 | 0.7218 | 0.7525 | 0.7526 | 0.7613 | 0.9833 | |
| | 0.0069 | 8.0 | 2104 | 0.0818 | 0.7921 | 0.7254 | 0.7573 | 0.7572 | 0.7670 | 0.9837 | |
| | 0.006 | 9.0 | 2367 | 0.0875 | 0.7624 | 0.7437 | 0.7529 | 0.7531 | 0.7587 | 0.9838 | |
| | 0.0048 | 10.0 | 2630 | 0.0919 | 0.7849 | 0.7126 | 0.7470 | 0.7471 | 0.7547 | 0.9831 | |
| | 0.0044 | 11.0 | 2893 | 0.0941 | 0.7755 | 0.7248 | 0.7493 | 0.7497 | 0.7613 | 0.9835 | |
| | 0.0035 | 12.0 | 3156 | 0.0996 | 0.7763 | 0.7431 | 0.7594 | 0.7596 | 0.7690 | 0.9838 | |
| | 0.0029 | 13.0 | 3419 | 0.0985 | 0.7498 | 0.7462 | 0.7480 | 0.7487 | 0.7596 | 0.9832 | |
| | 0.0028 | 14.0 | 3682 | 0.1044 | 0.7579 | 0.7010 | 0.7284 | 0.7285 | 0.7330 | 0.9827 | |
| | 0.0025 | 15.0 | 3945 | 0.1053 | 0.7613 | 0.7297 | 0.7452 | 0.7454 | 0.7517 | 0.9832 | |
| | 0.0024 | 16.0 | 4208 | 0.1024 | 0.7551 | 0.7395 | 0.7472 | 0.7474 | 0.7517 | 0.9834 | |
| | 0.0021 | 17.0 | 4471 | 0.1128 | 0.7843 | 0.7120 | 0.7464 | 0.7465 | 0.7579 | 0.9831 | |
| |
| |
| ### Framework versions |
| |
| - Transformers 4.48.2 |
| - Pytorch 2.4.1+cu121 |
| - Datasets 3.0.2 |
| - Tokenizers 0.21.0 |
| |
