Llama-3.1-8B-Instruct-EI1-2ep-sft-warmupbs
This model is a fine-tuned version of meta-llama/Llama-3.1-8B-Instruct on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6092
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: 6e-06
- train_batch_size: 1
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 16
- total_train_batch_size: 16
- total_eval_batch_size: 128
- optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.05
- num_epochs: 2.0
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| No log | 0.1403 | 100 | 0.6723 |
| No log | 0.2805 | 200 | 0.6387 |
| No log | 0.4208 | 300 | 0.6302 |
| No log | 0.5610 | 400 | 0.6238 |
| 0.6538 | 0.7013 | 500 | 0.6162 |
| 0.6538 | 0.8415 | 600 | 0.6104 |
| 0.6538 | 0.9818 | 700 | 0.6053 |
| 0.6538 | 1.1220 | 800 | 0.6186 |
| 0.6538 | 1.2623 | 900 | 0.6180 |
| 0.4905 | 1.4025 | 1000 | 0.6138 |
| 0.4905 | 1.5428 | 1100 | 0.6119 |
| 0.4905 | 1.6830 | 1200 | 0.6103 |
| 0.4905 | 1.8233 | 1300 | 0.6093 |
| 0.4905 | 1.9635 | 1400 | 0.6092 |
Framework versions
- Transformers 4.43.4
- Pytorch 2.4.0+cu121
- Datasets 3.0.1
- Tokenizers 0.19.1