CeLLaTe-ner-2class-tapt-pubmedbert-reinitllrd-tokenizer-adapted-combinedData-lr_3.215
This model is a fine-tuned version of Mardiyyah/CeLLaTe-tapt-pubmedbert-tokenizer-adapted-wwmask-combinedData on the OTAR3088/CeLLaTe-ner-2class-iob_final dataset. It achieves the following results on the evaluation set:
- Loss: 0.1216
- Precision: 0.7861
- Recall: 0.7601
- Micro F1: 0.7729
- Weighted F1: 0.7731
- Macro F1: 0.7810
- Accuracy: 0.9839
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: 3.215e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 3407
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- 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 |
|---|---|---|---|---|---|---|---|---|---|
| 0.1395 | 1.0 | 526 | 0.0609 | 0.7306 | 0.7468 | 0.7386 | 0.7388 | 0.7494 | 0.9829 |
| 0.0152 | 2.0 | 1052 | 0.0764 | 0.7473 | 0.7468 | 0.7471 | 0.7474 | 0.7607 | 0.9825 |
| 0.0072 | 3.0 | 1578 | 0.0989 | 0.7922 | 0.6831 | 0.7336 | 0.7334 | 0.7403 | 0.9825 |
| 0.0041 | 4.0 | 2104 | 0.0834 | 0.7472 | 0.7661 | 0.7565 | 0.7573 | 0.7684 | 0.9832 |
| 0.0027 | 5.0 | 2630 | 0.1005 | 0.7572 | 0.7595 | 0.7583 | 0.7588 | 0.7681 | 0.9828 |
| 0.0019 | 6.0 | 3156 | 0.1007 | 0.7334 | 0.7727 | 0.7526 | 0.7538 | 0.7659 | 0.9828 |
| 0.0015 | 7.0 | 3682 | 0.1114 | 0.8016 | 0.7360 | 0.7674 | 0.7674 | 0.7748 | 0.9841 |
| 0.0012 | 8.0 | 4208 | 0.1120 | 0.775 | 0.7643 | 0.7696 | 0.7698 | 0.7767 | 0.9837 |
| 0.0013 | 9.0 | 4734 | 0.1137 | 0.7363 | 0.7757 | 0.7555 | 0.7560 | 0.7662 | 0.9826 |
| 0.0012 | 10.0 | 5260 | 0.1213 | 0.7764 | 0.7685 | 0.7724 | 0.7728 | 0.7802 | 0.9840 |
| 0.0008 | 11.0 | 5786 | 0.1213 | 0.7861 | 0.7601 | 0.7729 | 0.7731 | 0.7810 | 0.9839 |
| 0.0007 | 12.0 | 6312 | 0.1268 | 0.7780 | 0.7565 | 0.7671 | 0.7671 | 0.7738 | 0.9838 |
| 0.0007 | 13.0 | 6838 | 0.1243 | 0.7806 | 0.7661 | 0.7733 | 0.7734 | 0.7798 | 0.9836 |
| 0.0005 | 14.0 | 7364 | 0.1237 | 0.7673 | 0.7673 | 0.7673 | 0.7676 | 0.7768 | 0.9835 |
Framework versions
- Transformers 4.48.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.2
- Tokenizers 0.21.0