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: 1.3342

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: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 3
  • total_train_batch_size: 24
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 2
  • num_epochs: 15
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
2.166 0.8571 2 1.8674
1.9742 1.7143 4 1.6570
1.0968 3.0 7 1.4801
1.5132 3.8571 9 1.4258
1.5467 4.7143 11 1.4036
0.9373 6.0 14 1.3799
1.3694 6.8571 16 1.3661
1.4186 7.7143 18 1.3553
0.9313 9.0 21 1.3434
1.322 9.8571 23 1.3380
1.3808 10.7143 25 1.3350
0.8632 12.0 28 1.3342
1.1499 12.8571 30 1.3342

Framework versions

  • PEFT 0.13.2
  • Transformers 4.45.2
  • Pytorch 2.4.1+cu121
  • Datasets 3.0.1
  • Tokenizers 0.20.1
Downloads last month
7
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no pipeline_tag.

Model tree for alekskusz/shawgpt-ft