CeLLaTe-tapt-bioformer16l-tokenizer-original_v2
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: 1.5001
- Accuracy: 0.7147
- Perplexity: 4.4820
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 |
|---|---|---|---|---|---|
| 1.698 | 1.0 | 15 | 1.5787 | 0.7114 | 4.8485 |
| 1.6839 | 2.0 | 30 | 1.5137 | 0.7147 | 4.5433 |
| 1.6846 | 3.0 | 45 | 1.5261 | 0.7109 | 4.6001 |
| 1.6437 | 4.0 | 60 | 1.4895 | 0.7200 | 4.4347 |
| 1.645 | 5.0 | 75 | 1.5032 | 0.7188 | 4.4961 |
| 1.5995 | 6.0 | 90 | 1.4965 | 0.7167 | 4.4660 |
| 1.5777 | 7.0 | 105 | 1.4347 | 0.7151 | 4.1985 |
| 1.5677 | 8.0 | 120 | 1.4652 | 0.7199 | 4.3282 |
| 1.5636 | 9.0 | 135 | 1.4912 | 0.7165 | 4.4422 |
| 1.5459 | 10.0 | 150 | 1.4509 | 0.7202 | 4.2671 |
| 1.5178 | 11.0 | 165 | 1.5107 | 0.7177 | 4.5301 |
| 1.5103 | 12.0 | 180 | 1.4657 | 0.7206 | 4.3307 |
Framework versions
- Transformers 4.48.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.2
- Tokenizers 0.21.0