GaetanMichelet's picture
End of training
37bce85 verified
metadata
base_model: mistralai/Mistral-7B-Instruct-v0.3
datasets:
  - GaetanMichelet/chat-60_ft_task-3_auto
library_name: peft
license: apache-2.0
tags:
  - alignment-handbook
  - trl
  - sft
  - generated_from_trainer
model-index:
  - name: Mistral-7B_task-3_60-samples_config-2_full_auto
    results: []

Mistral-7B_task-3_60-samples_config-2_full_auto

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on the GaetanMichelet/chat-60_ft_task-3_auto dataset. It achieves the following results on the evaluation set:

  • Loss: 0.8653

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
1.2142 0.6957 2 1.2086
1.2072 1.7391 5 1.1265
1.0872 2.7826 8 1.0536
1.0158 3.8261 11 0.9948
0.9378 4.8696 14 0.9304
0.8785 5.9130 17 0.8949
0.8173 6.9565 20 0.8793
0.8163 8.0 23 0.8703
0.7922 8.6957 25 0.8672
0.7598 9.7391 28 0.8669
0.7519 10.7826 31 0.8653
0.7182 11.8261 34 0.8703
0.6734 12.8696 37 0.8750
0.6525 13.9130 40 0.8917
0.6044 14.9565 43 0.8961
0.5698 16.0 46 0.9223
0.5067 16.6957 48 0.9411
0.51 17.7391 51 0.9559

Framework versions

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