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.

Sign up for GitHub

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

Jump to bottom

[Feature] Deprecate batch_size in favor of gradient accumulation steps #121

Closed
3 of 4 tasks
NanoCode012 opened this issue May 31, 2023 · 4 comments · Fixed by #130
Closed
3 of 4 tasks

[Feature] Deprecate batch_size in favor of gradient accumulation steps #121

NanoCode012 opened this issue May 31, 2023 · 4 comments · Fixed by #130
Labels
enhancement New feature or request

Comments

@NanoCode012
Copy link
Collaborator

NanoCode012 commented May 31, 2023 •
edited
Loading

Reason: Easier for calculation

  • Replace use of batch size with gradient
  • Replace in doc
  • Check duplicate config in validate_config
  • Replace all configs/examples
@NanoCode012 NanoCode012 added the enhancement New feature or request label May 31, 2023
@winglian
Copy link
Collaborator

sorry, didn't see this. temp fix for now #123

@NanoCode012
Copy link
Collaborator Author

I think your solution is quite nice! I however just noticed something

gas = cfg.batch_size // cfg.micro_batch_size

Should we have also divided this by # gpu?

@winglian
Copy link
Collaborator

that gets handled here:
https://github.com/OpenAccess-AI-Collective/axolotl/blob/main/scripts/finetune.py#L162-L166

@NanoCode012
Copy link
Collaborator Author

that gets handled here:
https://github.com/OpenAccess-AI-Collective/axolotl/blob/main/scripts/finetune.py#L162-L166

I see now!

@winglian winglian mentioned this issue May 31, 2023
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
enhancement New feature or request
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

swap batch size for gradient accumulation steps to decouple from num gpu axolotl-ai-cloud/axolotl
2 participants
@winglian @NanoCode012