CeLLaTe-ner-2class-bioformer16l-baseline-with-gazetteers-lr_2.10
This model is a fine-tuned version of bioformers/bioformer-16l on the OTAR3088/CeLLaTe-ner-2class-iob_final dataset. It achieves the following results on the evaluation set:
- Loss: 0.1077
- Precision: 0.7897
- Recall: 0.7607
- Micro F1: 0.7749
- Weighted F1: 0.7750
- Macro F1: 0.7861
- Accuracy: 0.9821
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: 2.1033235284928615e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 17
- 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: 0.02
- 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 |
|---|---|---|---|---|---|---|---|---|---|
| 0.158 | 1.0 | 1052 | 0.0647 | 0.7696 | 0.7379 | 0.7534 | 0.7538 | 0.7636 | 0.9810 |
| 0.02 | 2.0 | 2104 | 0.0788 | 0.7761 | 0.7209 | 0.7475 | 0.7480 | 0.7599 | 0.9804 |
| 0.01 | 3.0 | 3156 | 0.0915 | 0.7834 | 0.7585 | 0.7707 | 0.7707 | 0.7834 | 0.9816 |
| 0.0059 | 4.0 | 4208 | 0.0911 | 0.7831 | 0.7652 | 0.7740 | 0.7742 | 0.7831 | 0.9824 |
| 0.0037 | 5.0 | 5260 | 0.1031 | 0.7796 | 0.7513 | 0.7652 | 0.7654 | 0.7805 | 0.9818 |
| 0.0026 | 6.0 | 6312 | 0.1077 | 0.7897 | 0.7607 | 0.7749 | 0.7750 | 0.7861 | 0.9821 |
| 0.002 | 7.0 | 7364 | 0.1238 | 0.8114 | 0.7388 | 0.7734 | 0.7732 | 0.7846 | 0.9820 |
| 0.0017 | 8.0 | 8416 | 0.1198 | 0.7840 | 0.7581 | 0.7708 | 0.7714 | 0.7832 | 0.9817 |
| 0.0014 | 9.0 | 9468 | 0.1343 | 0.7722 | 0.7245 | 0.7476 | 0.7478 | 0.7589 | 0.9809 |
Framework versions
- Transformers 4.48.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.2
- Tokenizers 0.21.0