Hugging Face's logo

hussenmi
/
fungpt-ft

PEFT
TensorBoard
Safetensors
Generated from Trainer
Model card Files Metrics Community
fungpt-ft / README.md
hussenmi's picture
hussenmi
hussenmi/fungpt-ft
7a5dcfb verified
preview code
|
raw
history blame
1.96 kB
metadata
license: apache-2.0
library_name: peft
tags:
  - generated_from_trainer
base_model: TheBloke/Mistral-7B-Instruct-v0.2-GPTQ
model-index:
  - name: fungpt-ft
    results: []

fungpt-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: 3.1043

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
5.2519 1.0 15 3.8834
3.6084 2.0 30 3.4631
3.1807 3.0 45 3.1895
2.921 4.0 60 3.1170
2.8287 5.0 75 3.0935
2.7328 6.0 90 3.0807
2.6611 7.0 105 3.0841
2.6403 8.0 120 3.0969
2.5925 9.0 135 3.1010
2.55 10.0 150 3.1043

Framework versions

  • PEFT 0.9.0
  • Transformers 4.38.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2