GaetanMichelet's picture
End of training
778e2a0 verified
metadata
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
datasets:
  - GaetanMichelet/chat-60_ft_task-2
library_name: peft
license: llama3.1
tags:
  - alignment-handbook
  - trl
  - sft
  - generated_from_trainer
model-index:
  - name: Llama-31-8B_task-2_60-samples_config-2_full
    results: []

Llama-31-8B_task-2_60-samples_config-2_full

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the GaetanMichelet/chat-60_ft_task-2 dataset. It achieves the following results on the evaluation set:

  • Loss: 1.0628

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
1.5658 0.6957 2 1.5799
1.5614 1.7391 5 1.5293
1.4751 2.7826 8 1.4358
1.3878 3.8261 11 1.3590
1.3039 4.8696 14 1.2951
1.2412 5.9130 17 1.2337
1.1483 6.9565 20 1.1724
1.1126 8.0 23 1.1255
1.0651 8.6957 25 1.1110
1.0248 9.7391 28 1.0985
1.0323 10.7826 31 1.0890
1.0119 11.8261 34 1.0821
0.9765 12.8696 37 1.0758
0.9796 13.9130 40 1.0704
0.9428 14.9565 43 1.0683
0.9511 16.0 46 1.0637
0.9117 16.6957 48 1.0628
0.9208 17.7391 51 1.0629
0.9108 18.7826 54 1.0634
0.8903 19.8261 57 1.0640
0.8762 20.8696 60 1.0664
0.8702 21.9130 63 1.0685
0.8465 22.9565 66 1.0700
0.8409 24.0 69 1.0734

Framework versions

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