BC5CDR_bioBERT_NER / README.md
judithrosell's picture
End of training
1706228
metadata
base_model: dmis-lab/biobert-v1.1
tags:
  - generated_from_trainer
model-index:
  - name: BC5CDR_bioBERT_NER
    results: []

BC5CDR_bioBERT_NER

This model is a fine-tuned version of dmis-lab/biobert-v1.1 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0808

  • Seqeval classification report: precision recall f1-score support

    Chemical 0.99 0.98 0.98 103336 Disease 0.83 0.88 0.85 6944

    micro avg 0.97 0.97 0.97 110280 macro avg 0.91 0.93 0.92 110280

weighted avg 0.98 0.97 0.97 110280

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: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 32
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 3

Training results

Training Loss Epoch Step Validation Loss Seqeval classification report
No log 1.0 143 0.0903 precision recall f1-score support
Chemical       0.98      0.98      0.98    103336
 Disease       0.79      0.87      0.83      6944

micro avg 0.97 0.97 0.97 110280 macro avg 0.89 0.92 0.90 110280 weighted avg 0.97 0.97 0.97 110280 | | No log | 2.0 | 286 | 0.0823 | precision recall f1-score support

Chemical       0.99      0.98      0.98    103336
 Disease       0.79      0.87      0.83      6944

micro avg 0.97 0.97 0.97 110280 macro avg 0.89 0.92 0.91 110280 weighted avg 0.97 0.97 0.97 110280 | | No log | 3.0 | 429 | 0.0808 | precision recall f1-score support

Chemical       0.99      0.98      0.98    103336
 Disease       0.83      0.88      0.85      6944

micro avg 0.97 0.97 0.97 110280 macro avg 0.91 0.93 0.92 110280 weighted avg 0.98 0.97 0.97 110280 |

Framework versions

  • Transformers 4.35.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.15.0
  • Tokenizers 0.15.0