GaetanMichelet's picture
End of training
184a880 verified
metadata
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
datasets:
  - GaetanMichelet/chat-60_ft_task-3
  - GaetanMichelet/chat-120_ft_task-3
  - GaetanMichelet/chat-180_ft_task-3
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-1_full
    results: []

Llama-31-8B_task-3_180-samples_config-1_full

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the GaetanMichelet/chat-60_ft_task-3, the GaetanMichelet/chat-120_ft_task-3 and the GaetanMichelet/chat-180_ft_task-3 datasets. It achieves the following results on the evaluation set:

  • Loss: 1.1361

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
1.6392 1.0 17 1.6279
1.4618 2.0 34 1.4343
1.2006 3.0 51 1.2241
1.0799 4.0 68 1.1761
1.0615 5.0 85 1.1524
1.0045 6.0 102 1.1361
0.9831 7.0 119 1.1392
0.8698 8.0 136 1.1567
0.7759 9.0 153 1.1918
0.7296 10.0 170 1.2537
0.6747 11.0 187 1.2852
0.4777 12.0 204 1.3877
0.5052 13.0 221 1.4002

Framework versions

  • PEFT 0.12.0
  • Transformers 4.44.0
  • Pytorch 2.1.2+cu121
  • Datasets 2.20.0
  • Tokenizers 0.19.1