CeLLaTe-tapt-bioformer16l-tokenizer-original-baseline_no_groupText
This model is a fine-tuned version of bioformers/bioformer-16l on the Mardiyyah/TAPT_CeLLaTe2.0 dataset. It achieves the following results on the evaluation set:
- Loss: 1.3755
- Accuracy: 0.7221
- Perplexity: 3.9572
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 |
|---|---|---|---|---|---|
| 1.162 | 1.0 | 1 | 1.2860 | 0.7353 | 3.6182 |
| 1.2649 | 2.0 | 2 | 1.2482 | 0.7288 | 3.4841 |
| 1.2806 | 3.0 | 3 | 1.2366 | 0.7260 | 3.4439 |
| 1.2221 | 4.0 | 4 | 1.1888 | 0.7578 | 3.2832 |
| 1.3136 | 5.0 | 5 | 1.1949 | 0.7511 | 3.3033 |
| 1.3282 | 6.0 | 6 | 1.2321 | 0.7289 | 3.4284 |
| 1.283 | 7.0 | 7 | 1.2136 | 0.7157 | 3.3656 |
| 1.2282 | 8.0 | 8 | 1.1614 | 0.7404 | 3.1945 |
| 1.3113 | 9.0 | 9 | 1.1683 | 0.7448 | 3.2164 |
| 1.1744 | 10.0 | 10 | 1.1764 | 0.7433 | 3.2425 |
| 1.1868 | 11.0 | 11 | 1.1305 | 0.7580 | 3.0972 |
| 1.2028 | 12.0 | 12 | 1.3218 | 0.7182 | 3.7503 |
| 1.1303 | 13.0 | 13 | 1.4027 | 0.6992 | 4.0663 |
| 1.2429 | 14.0 | 14 | 1.2150 | 0.7514 | 3.3701 |
| 1.2672 | 15.0 | 15 | 1.2561 | 0.7347 | 3.5118 |
| 1.1134 | 16.0 | 16 | 1.3824 | 0.7135 | 3.9846 |
Framework versions
- Transformers 4.48.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.2
- Tokenizers 0.21.0