Biobert-base-cased-v1.2-finetuned-ner-CRAFT
This model is a fine-tuned version of dmis-lab/biobert-base-cased-v1.2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.1878
- Precision: 0.8397
- Recall: 0.8366
- F1: 0.8382
- Accuracy: 0.9683
Model description
This model performs Named Entity Recognition for 6 entity tags: Sequence, Cell, Protein, Gene, Taxon, and Chemical from the CRAFT(Colorado Richly Annotated Full Text) Corpus in English. Entity tags have been normalized and replaced from the original three letter code to a full name e.g. B-Protein, I-Chemical.
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: 3e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.11 | 1.0 | 1360 | 0.1668 | 0.7952 | 0.7917 | 0.7934 | 0.9611 |
| 0.0484 | 2.0 | 2720 | 0.1640 | 0.8224 | 0.8371 | 0.8297 | 0.9661 |
| 0.0261 | 3.0 | 4080 | 0.1812 | 0.8143 | 0.8447 | 0.8292 | 0.9662 |
| 0.0112 | 4.0 | 5440 | 0.1878 | 0.8397 | 0.8366 | 0.8382 | 0.9683 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu111
- Datasets 1.18.4
- Tokenizers 0.11.6