CeLLaTe-tapt-bioformer16l-tokenizer-original-baseline-combinedData
This model is a fine-tuned version of bioformers/bioformer-16l on the Mardiyyah/Combined_TAPT_dataset dataset. It achieves the following results on the evaluation set:
- Loss: 1.1916
- Accuracy: 0.7443
- Perplexity: 3.2923
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: 3e-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-06 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.06
- num_epochs: 100
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Perplexity |
|---|---|---|---|---|---|
| 1.4889 | 1.0 | 29 | 1.2740 | 0.7351 | 3.5751 |
| 1.4585 | 2.0 | 58 | 1.2478 | 0.7370 | 3.4827 |
| 1.4285 | 3.0 | 87 | 1.2533 | 0.7370 | 3.5018 |
| 1.4172 | 4.0 | 116 | 1.2615 | 0.7337 | 3.5307 |
| 1.418 | 5.0 | 145 | 1.2442 | 0.7353 | 3.4701 |
| 1.377 | 6.0 | 174 | 1.2396 | 0.7365 | 3.4542 |
| 1.3549 | 7.0 | 203 | 1.2129 | 0.7407 | 3.3632 |
| 1.3621 | 8.0 | 232 | 1.2367 | 0.7366 | 3.4444 |
| 1.3395 | 9.0 | 261 | 1.2343 | 0.7358 | 3.4360 |
| 1.3192 | 10.0 | 290 | 1.1958 | 0.7452 | 3.3062 |
| 1.3043 | 11.0 | 319 | 1.2124 | 0.7431 | 3.3615 |
| 1.2942 | 12.0 | 348 | 1.2456 | 0.7354 | 3.4751 |
| 1.2804 | 13.0 | 377 | 1.2211 | 0.7399 | 3.3910 |
| 1.2711 | 14.0 | 406 | 1.1851 | 0.7448 | 3.2709 |
| 1.2521 | 15.0 | 435 | 1.2367 | 0.7371 | 3.4443 |
| 1.2576 | 16.0 | 464 | 1.2002 | 0.7417 | 3.3206 |
| 1.2496 | 17.0 | 493 | 1.2007 | 0.7409 | 3.3226 |
| 1.2302 | 18.0 | 522 | 1.2134 | 0.7399 | 3.3649 |
| 1.2186 | 19.0 | 551 | 1.2047 | 0.7430 | 3.3359 |
Framework versions
- Transformers 4.48.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.2
- Tokenizers 0.21.0