chansung's picture
End of training
d24b7d3 verified
|
raw
history blame
2.03 kB
metadata
license: gemma
library_name: peft
tags:
  - alignment-handbook
  - trl
  - sft
  - generated_from_trainer
base_model: google/gemma-7b
datasets:
  - llama-duo/synth_summarize_dataset
model-index:
  - name: gemma7b-summarize-claude3sonnet-30k
    results: []

Visualize in Weights & Biases

gemma7b-summarize-claude3sonnet-30k

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

  • Loss: 2.3268

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: 2
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 2
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 16
  • total_eval_batch_size: 4
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 5

Training results

Training Loss Epoch Step Validation Loss
0.7892 1.0 205 2.2406
0.7033 2.0 410 2.2004
0.6346 3.0 615 2.2458
0.5755 4.0 820 2.2935
0.5144 5.0 1025 2.3268

Framework versions

  • PEFT 0.11.1
  • Transformers 4.41.0
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1