chansung's picture
End of training
a598734 verified
metadata
license: gemma
library_name: peft
tags:
  - alignment-handbook
  - trl
  - sft
  - generated_from_trainer
base_model: google/gemma-2b
datasets:
  - llama-duo/synth_summarize_dataset_dedup
model-index:
  - name: gemma2b-summarize-gpt4o-2k
    results: []

gemma2b-summarize-gpt4o-2k

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

  • Loss: 2.5878

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: 8
  • eval_batch_size: 8
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 3
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 48
  • total_eval_batch_size: 24
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss
2.9978 1.0 5 3.1071
2.5123 2.0 10 2.8503
2.2077 3.0 15 2.7154
1.9749 4.0 20 2.6507
1.8015 5.0 25 2.6242
1.6817 6.0 30 2.6105
1.6095 7.0 35 2.6003
1.5701 8.0 40 2.5917
1.5524 9.0 45 2.5882
1.5443 10.0 50 2.5878

Framework versions

  • PEFT 0.11.1
  • Transformers 4.41.2
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.2
  • Tokenizers 0.19.1