Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[チューニング] - nemo-aligner での DPO 学習 #91

Open
Taka008 opened this issue Nov 29, 2024 · 0 comments
Open

[チューニング] - nemo-aligner での DPO 学習 #91

Taka008 opened this issue Nov 29, 2024 · 0 comments
Assignees
Labels
tuning Experiment of model fine-tuning

Comments

@Taka008
Copy link

Taka008 commented Nov 29, 2024

Overview

llm-jp-nemo-aligner で DPO の学習を行えるようにする

Details

llm-jp-nemo-aligner で DPO の学習を行えるように学習スクリプトを整備する

llm-jp-3-1.8b or llm-jp-3-13b を weblab-GENIAC/aya-ja-evol-instruct-calm3-dpo-masked で学習させる

Resources

  • 計算機
    • クラスタ: llm-jp-nvlink
    • ノード種別: gpu
    • ノード台数: 4
  • コード
  • 入力データ:
    • {name}: {physical path}
  • 出力データ:
    • 保存先: {cluster}:/data/experiments/{number}
    • データ内訳:
      • {name}: xxx TB (バッファ容量を含む)
  • W&B ログ:
  • 開始日: 2024-11-29
  • 終了予定日: 2024-12-31
@Taka008 Taka008 added pretrain Experiment of model pretrain tuning Experiment of model fine-tuning and removed pretrain Experiment of model pretrain labels Nov 29, 2024
@Taka008 Taka008 self-assigned this Nov 29, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
tuning Experiment of model fine-tuning
Projects
None yet
Development

No branches or pull requests

1 participant