This model is a fine-tuned version of Mardiyyah/CeLLaTe-tapt-pubmedbert-tokenizer-adapted-wholewordmask on the OTAR3088/CeLLaTe-ner-3class-iob_final dataset. It achieves the following results on the evaluation set:
More information needed
More information needed
More information needed
The following hyperparameters were used during training:
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | Micro F1 | Weighted F1 | Macro F1 | Accuracy |
|---|---|---|---|---|---|---|---|---|---|
| 0.3718 | 1.0 | 263 | 0.0794 | 0.5110 | 0.5659 | 0.5371 | 0.5271 | 0.5014 | 0.9751 |
| 0.058 | 2.0 | 526 | 0.0637 | 0.7644 | 0.7111 | 0.7368 | 0.7362 | 0.7354 | 0.9815 |
| 0.0355 | 3.0 | 789 | 0.0709 | 0.6875 | 0.7387 | 0.7122 | 0.7136 | 0.7147 | 0.9795 |
| 0.0254 | 4.0 | 1052 | 0.0747 | 0.7619 | 0.7194 | 0.7400 | 0.7391 | 0.7365 | 0.9818 |
| 0.0185 | 5.0 | 1315 | 0.0762 | 0.7787 | 0.7033 | 0.7391 | 0.7394 | 0.7382 | 0.9816 |
| 0.0152 | 6.0 | 1578 | 0.0899 | 0.7295 | 0.7215 | 0.7255 | 0.7261 | 0.7255 | 0.9807 |
| 0.0111 | 7.0 | 1841 | 0.0888 | 0.7619 | 0.7361 | 0.7487 | 0.7478 | 0.7451 | 0.9823 |
| 0.0088 | 8.0 | 2104 | 0.0928 | 0.7398 | 0.7251 | 0.7324 | 0.7322 | 0.7313 | 0.9814 |
| 0.0077 | 9.0 | 2367 | 0.0982 | 0.7728 | 0.7470 | 0.7597 | 0.7580 | 0.7545 | 0.9827 |
| 0.0067 | 10.0 | 2630 | 0.1019 | 0.7891 | 0.7205 | 0.7532 | 0.7521 | 0.7508 | 0.9826 |
| 0.0058 | 11.0 | 2893 | 0.1009 | 0.7676 | 0.7512 | 0.7593 | 0.7583 | 0.7559 | 0.9828 |
| 0.005 | 12.0 | 3156 | 0.0999 | 0.7545 | 0.7569 | 0.7557 | 0.7555 | 0.7537 | 0.9830 |
| 0.0043 | 13.0 | 3419 | 0.1139 | 0.7456 | 0.7507 | 0.7481 | 0.7475 | 0.7469 | 0.9822 |
| 0.004 | 14.0 | 3682 | 0.1113 | 0.7745 | 0.7277 | 0.7504 | 0.7498 | 0.7471 | 0.9820 |
| 0.0033 | 15.0 | 3945 | 0.1086 | 0.7469 | 0.7605 | 0.7537 | 0.7539 | 0.7525 | 0.9821 |
| 0.0029 | 16.0 | 4208 | 0.1117 | 0.7648 | 0.7480 | 0.7563 | 0.7561 | 0.7551 | 0.9825 |
model.safetensors 436 MBtraining_args.bin 5.8 kB