metadata
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
datasets:
- GaetanMichelet/chat-60_ft_task-3_auto
- GaetanMichelet/chat-120_ft_task-3_auto
- GaetanMichelet/chat-180_ft_task-3_auto
library_name: peft
license: llama3.1
tags:
- alignment-handbook
- trl
- sft
- generated_from_trainer
model-index:
- name: Llama-31-8B_task-3_180-samples_config-2_auto
results: []
Llama-31-8B_task-3_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-3_auto, the GaetanMichelet/chat-120_ft_task-3_auto and the GaetanMichelet/chat-180_ft_task-3_auto datasets. It achieves the following results on the evaluation set:
- Loss: 0.3387
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 |
|---|---|---|---|
| 2.1694 | 0.9412 | 8 | 2.0003 |
| 1.1443 | 2.0 | 17 | 0.8652 |
| 0.5201 | 2.9412 | 25 | 0.4725 |
| 0.2979 | 4.0 | 34 | 0.3815 |
| 0.3128 | 4.9412 | 42 | 0.3738 |
| 0.3107 | 6.0 | 51 | 0.3410 |
| 0.2488 | 6.9412 | 59 | 0.3388 |
| 0.252 | 8.0 | 68 | 0.3387 |
| 0.2324 | 8.9412 | 76 | 0.3474 |
| 0.1655 | 10.0 | 85 | 0.3932 |
| 0.1782 | 10.9412 | 93 | 0.4127 |
| 0.0536 | 12.0 | 102 | 0.4492 |
| 0.0425 | 12.9412 | 110 | 0.5433 |
| 0.0347 | 14.0 | 119 | 0.6090 |
| 0.0231 | 14.9412 | 127 | 0.5974 |
Framework versions
- PEFT 0.12.0
- Transformers 4.44.0
- Pytorch 2.1.2+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1