Skip to content

Minor fix: ZeRO-1 grad clipping#4796

Merged
alanwaketan merged 2 commits intopytorch:masterfrom
hgt312:fix_zero1
May 2, 2023
Merged

Minor fix: ZeRO-1 grad clipping#4796
alanwaketan merged 2 commits intopytorch:masterfrom
hgt312:fix_zero1

Conversation

@hgt312
Copy link
Copy Markdown
Collaborator

@hgt312 hgt312 commented Mar 17, 2023

implementation and test in previous PR #4648

reduce local norm across shards

@hgt312
Copy link
Copy Markdown
Collaborator Author

hgt312 commented Mar 20, 2023

@JackCaoG could you help add some reviewers?

@JackCaoG
Copy link
Copy Markdown
Collaborator

@hgt312 sorry I was out for last few weeks, let me try to take a look.

@alanwaketan
Copy link
Copy Markdown
Collaborator

Do we have a test case already?

@hgt312
Copy link
Copy Markdown
Collaborator Author

hgt312 commented Apr 15, 2023

Do we have a test case already?

yes, in #4648

@alanwaketan
Copy link
Copy Markdown
Collaborator

Do we have a test case already?

yes, in #4648

Thanks. Can you mention that in the description?

Copy link
Copy Markdown
Collaborator

@alanwaketan alanwaketan left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM.

@alanwaketan alanwaketan merged commit 71f9a35 into pytorch:master May 2, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants