Hugging Face's logo

Baidicoot
/
run-gemma

PEFT
Safetensors
trl
reward-trainer
Generated from Trainer
Model card Files Community
run-gemma / README.md
Baidicoot's picture
Baidicoot
End of training
054f29d verified
preview code
|
raw
history blame
2.59 kB
metadata
base_model: google/gemma-2b
library_name: peft
license: gemma
metrics:
  - accuracy
tags:
  - trl
  - reward-trainer
  - generated_from_trainer
model-index:
  - name: run-gemma
    results: []

run-gemma

This model is a fine-tuned version of google/gemma-2b on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4736
  • Accuracy: 0.8

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: 2e-05
  • train_batch_size: 64
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • training_steps: 200

Training results

Training Loss Epoch Step Validation Loss Accuracy
0.7109 0.2941 10 0.7341 0.3684
0.7148 0.5882 20 0.7228 0.3947
0.6992 0.8824 30 0.7039 0.4895
0.668 1.1765 40 0.6780 0.5789
0.6328 1.4706 50 0.6501 0.6474
0.6211 1.7647 60 0.6144 0.6737
0.5781 2.0588 70 0.5833 0.6842
0.5156 2.3529 80 0.5591 0.7
0.5352 2.6471 90 0.5355 0.7368
0.5508 2.9412 100 0.5179 0.7421
0.543 3.2353 110 0.4917 0.7579
0.4141 3.5294 120 0.4833 0.7684
0.4102 3.8235 130 0.4706 0.7737
0.4316 4.1176 140 0.4643 0.7789
0.4844 4.4118 150 0.4683 0.8
0.4199 4.7059 160 0.4668 0.8
0.4082 5.0 170 0.4795 0.7842
0.3516 5.2941 180 0.4804 0.7842
0.4238 5.5882 190 0.4826 0.7947
0.3867 5.8824 200 0.4736 0.8

Framework versions

  • PEFT 0.12.0
  • Transformers 4.44.0
  • Pytorch 2.4.0+cu121
  • Datasets 2.20.0
  • Tokenizers 0.19.1