metadata
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
datasets:
- GaetanMichelet/chat-60_ft_task-1
- GaetanMichelet/chat-120_ft_task-1
- GaetanMichelet/chat-180_ft_task-1
library_name: peft
license: llama3.1
tags:
- alignment-handbook
- trl
- sft
- generated_from_trainer
model-index:
- name: Llama-31-8B_task-1_180-samples_config-1
results: []
Llama-31-8B_task-1_180-samples_config-1
This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the GaetanMichelet/chat-60_ft_task-1, the GaetanMichelet/chat-120_ft_task-1 and the GaetanMichelet/chat-180_ft_task-1 datasets. It achieves the following results on the evaluation set:
- Loss: 1.2264
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: 8
- total_train_batch_size: 8
- 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 |
|---|---|---|---|
| 1.8393 | 1.0 | 17 | 1.7238 |
| 1.3906 | 2.0 | 34 | 1.4334 |
| 1.1829 | 3.0 | 51 | 1.2720 |
| 1.0037 | 4.0 | 68 | 1.2264 |
| 0.875 | 5.0 | 85 | 1.2729 |
| 0.5286 | 6.0 | 102 | 1.4641 |
| 0.2721 | 7.0 | 119 | 1.7061 |
| 0.16 | 8.0 | 136 | 2.0944 |
| 0.0814 | 9.0 | 153 | 2.3201 |
| 0.0608 | 10.0 | 170 | 2.2903 |
| 0.033 | 11.0 | 187 | 2.6819 |
Framework versions
- PEFT 0.12.0
- Transformers 4.44.0
- Pytorch 2.1.2+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1