CeLLaTe-tapt-bioformer16l-tokenizer-original-wwmask-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: 1.3169
- Accuracy: 0.7273
- Perplexity: 3.7320
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 |
|---|---|---|---|---|---|
| 1.5813 | 1.0 | 29 | 1.3846 | 0.7146 | 3.9930 |
| 1.5702 | 2.0 | 58 | 1.3532 | 0.7202 | 3.8697 |
| 1.5464 | 3.0 | 87 | 1.3704 | 0.7165 | 3.9369 |
| 1.5274 | 4.0 | 116 | 1.3697 | 0.7177 | 3.9340 |
| 1.51 | 5.0 | 145 | 1.3292 | 0.7220 | 3.7779 |
| 1.5048 | 6.0 | 174 | 1.3540 | 0.7222 | 3.8730 |
| 1.481 | 7.0 | 203 | 1.3357 | 0.7218 | 3.8025 |
| 1.4711 | 8.0 | 232 | 1.3570 | 0.7214 | 3.8844 |
| 1.4585 | 9.0 | 261 | 1.3379 | 0.7252 | 3.8111 |
| 1.452 | 10.0 | 290 | 1.3466 | 0.7221 | 3.8444 |
Framework versions
- Transformers 4.48.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.2
- Tokenizers 0.21.0