Hugging Face's logo

himanshue2e
/
gemma-2b-gemma-finetune

PEFT
Safetensors
trl
sft
Generated from Trainer
Model card Files Community
gemma-2b-gemma-finetune / README.md
Himanshu singh
End of training
bca0a66 verified
preview code
|
raw
history blame
1.64 kB
metadata
license: other
library_name: peft
tags:
  - trl
  - sft
  - generated_from_trainer
base_model: google/gemma-2b
datasets:
  - generator
model-index:
  - name: gemma-2b-gemma-finetune
    results: []

gemma-2b-gemma-finetune

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

  • Loss: 1.4721

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: 2.5e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 32
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • training_steps: 10

Training results

Training Loss Epoch Step Validation Loss
No log 0.1 2 1.5211
No log 0.2 4 1.4995
1.4799 0.3 6 1.4839
1.4799 0.41 8 1.4758
1.4367 0.51 10 1.4721

Framework versions

  • PEFT 0.10.1.dev0
  • Transformers 4.40.0.dev0
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2