GaetanMichelet's picture
End of training
ac24265 verified
metadata
base_model: mistralai/Mistral-7B-Instruct-v0.3
datasets:
  - GaetanMichelet/chat-60_ft_task-2
library_name: peft
license: apache-2.0
tags:
  - alignment-handbook
  - trl
  - sft
  - generated_from_trainer
model-index:
  - name: Mistral-7B_task-2_60-samples_config-2
    results: []

Mistral-7B_task-2_60-samples_config-2

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on the GaetanMichelet/chat-60_ft_task-2 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5370

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
0.9782 0.6957 2 0.9433
0.9325 1.7391 5 0.7483
0.657 2.7826 8 0.6180
0.5682 3.8261 11 0.5942
0.5483 4.8696 14 0.5666
0.4875 5.9130 17 0.5441
0.4347 6.9565 20 0.5370
0.3995 8.0 23 0.5381
0.3709 8.6957 25 0.5431
0.2801 9.7391 28 0.5705
0.2582 10.7826 31 0.6300
0.1682 11.8261 34 0.6605
0.1192 12.8696 37 0.7693
0.0549 13.9130 40 0.8641

Framework versions

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