Skip to content

Conversation

@LorrinWWW
Copy link
Contributor

This PR did several things:

  • Add feedback data in data/OIG/prepare.py
  • Add a fine-tuning script in training/finetune_Pythia-Chat-Base-7B-feedback.sh, which further fine-tune upon the ckpt produced by training/finetune_Pythia-Chat-Base-7B.sh.
  • Some trivial changes:
    • Add --checkpoint-load-path: load another ckpt before training starts
    • Restart step counting with --init-steps

@LorrinWWW LorrinWWW requested a review from csris March 31, 2023 03:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant