Llama-31-8B_task-1_180-samples_config-2_auto
This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the GaetanMichelet/chat-60_ft_task-1_auto, the GaetanMichelet/chat-120_ft_task-1_auto and the GaetanMichelet/chat-180_ft_task-1_auto datasets.
It achieves the following results on the evaluation set:
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: 0.0001
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 16
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
Training results
| Training Loss |
Epoch |
Step |
Validation Loss |
| 1.178 |
0.9412 |
8 |
1.1844 |
| 0.9678 |
2.0 |
17 |
1.0291 |
| 0.9547 |
2.9412 |
25 |
0.9495 |
| 0.8037 |
4.0 |
34 |
0.8970 |
| 0.7404 |
4.9412 |
42 |
0.8755 |
| 0.6681 |
6.0 |
51 |
0.9058 |
| 0.4752 |
6.9412 |
59 |
0.9785 |
| 0.3663 |
8.0 |
68 |
1.0201 |
| 0.2328 |
8.9412 |
76 |
1.2509 |
| 0.1375 |
10.0 |
85 |
1.4120 |
| 0.1013 |
10.9412 |
93 |
1.4669 |
| 0.0523 |
12.0 |
102 |
1.5482 |
Framework versions
- PEFT 0.12.0
- Transformers 4.44.0
- Pytorch 2.1.2+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1
