yimingzhang's picture
End of training
1d19bad verified
metadata
license: gemma
base_model: google/gemma-7b
tags:
  - alignment-handbook
  - trl
  - sft
  - generated_from_trainer
  - trl
  - sft
  - alignment-handbook
  - generated_from_trainer
datasets:
  - yimingzhang/backtrack-0522
model-index:
  - name: gemma-backtrack-0522
    results: []

Visualize in Weights & Biases

gemma-backtrack-0522

This model is a fine-tuned version of google/gemma-7b on the yimingzhang/backtrack-0522 dataset. It achieves the following results on the evaluation set:

  • Loss: 24.1739

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: 4
  • eval_batch_size: 4
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 8
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 128
  • total_eval_batch_size: 32
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 3

Training results

Training Loss Epoch Step Validation Loss
63.281 0.8571 3 31.5266
53.4745 2.0 7 24.9424
53.4745 2.5714 9 24.1739

Framework versions

  • Transformers 4.41.0
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1