shawgpt-ft / README.md
shawhin's picture
shawhin/shawgpt-ft
af78907 verified
metadata
license: apache-2.0
base_model: TheBloke/Mistral-7B-Instruct-v0.2-GPTQ
tags:
  - generated_from_trainer
model-index:
  - name: shawgpt-ft
    results: []

shawgpt-ft

This model is a fine-tuned version of TheBloke/Mistral-7B-Instruct-v0.2-GPTQ on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 1.7578

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: 0.0002
  • train_batch_size: 4
  • eval_batch_size: 4
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 2
  • num_epochs: 10
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
4.5944 0.92 3 3.9680
4.0632 1.85 6 3.4430
3.4599 2.77 9 2.9820
2.2563 4.0 13 2.5456
2.6364 4.92 16 2.2781
2.2961 5.85 19 2.0657
2.0692 6.77 22 1.9161
1.4295 8.0 26 1.8044
1.8307 8.92 29 1.7627
1.2551 9.23 30 1.7578

Framework versions

  • Transformers 4.35.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.17.1
  • Tokenizers 0.15.2