babylama-activation_gelu
This model is a fine-tuned version of Harshatheeswar/babylama-attention_heads12 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 3.8124
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: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 64
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 4.0596 | 0.9999 | 5559 | 4.1325 |
| 3.7904 | 1.9999 | 11119 | 3.8984 |
| 3.598 | 2.9998 | 16678 | 3.8069 |
| 3.3921 | 3.9999 | 22238 | 3.7945 |
| 3.3118 | 4.9994 | 27795 | 3.8124 |
Framework versions
- Transformers 4.46.2
- Pytorch 2.5.0+cu121
- Datasets 3.1.0
- Tokenizers 0.20.3