Llama-450M
This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 5.8986
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: 0.0003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 50
- num_epochs: 15
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 8.6051 | 0.89 | 2 | 8.5427 |
| 8.1233 | 1.78 | 4 | 8.2081 |
| 7.2688 | 2.67 | 6 | 7.6786 |
| 6.3982 | 4.0 | 9 | 7.0782 |
| 5.8794 | 4.89 | 11 | 6.7779 |
| 5.4786 | 5.78 | 13 | 6.5717 |
| 4.994 | 6.67 | 15 | 6.3356 |
| 4.35 | 8.0 | 18 | 6.2257 |
| 3.9757 | 8.89 | 20 | 6.0451 |
| 3.4479 | 9.78 | 22 | 6.0242 |
| 3.1004 | 10.67 | 24 | 5.9219 |
| 2.5207 | 12.0 | 27 | 5.8224 |
| 2.1123 | 12.89 | 29 | 5.9286 |
| 1.7641 | 13.33 | 30 | 5.8986 |
Framework versions
- Transformers 4.39.1
- Pytorch 2.1.2+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0