-
Notifications
You must be signed in to change notification settings - Fork 3.4k
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
[FIX] ShardedGradScaler should only be set for FP16 #12915
Merged
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
SeanNaren
requested review from
tchaton,
awaelchli,
justusschock,
williamFalcon,
Borda,
carmocca,
kaushikb11 and
rohitgr7
as code owners
April 28, 2022 10:14
carmocca
approved these changes
Apr 28, 2022
carmocca
added
bug
Something isn't working
strategy: fairscale sharded (removed)
Sharded Data Parallel
precision: amp
Automatic Mixed Precision
labels
Apr 28, 2022
kaushikb11
approved these changes
Apr 28, 2022
justusschock
approved these changes
Apr 28, 2022
carmocca
reviewed
Apr 28, 2022
Co-authored-by: Carlos Mocholí <[email protected]>
otaj
approved these changes
Apr 28, 2022
rohitgr7
reviewed
Apr 28, 2022
rohitgr7
approved these changes
Apr 28, 2022
carmocca
added a commit
that referenced
this pull request
Apr 28, 2022
Co-authored-by: Carlos Mocholí <[email protected]>
lexierule
pushed a commit
that referenced
this pull request
May 3, 2022
Co-authored-by: Carlos Mocholí <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Labels
bug
Something isn't working
precision: amp
Automatic Mixed Precision
ready
PRs ready to be merged
strategy: fairscale sharded (removed)
Sharded Data Parallel
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
Closes #12914
When running with sharded training + bf16, I noticed that the ShardedGradScaler was being passed. This is incorrect as BF16 does not require any scaling logic. Fixed this and added a unit test!
Before submitting
PR review
Anyone in the community is welcome to review the PR.
Before you start reviewing, make sure you have read the review guidelines. In short, see the following bullet-list:
Did you have fun?
Make sure you had fun coding 🙃
cc @Borda @SeanNaren @awaelchli @rohitgr7 @akihironitta @carmocca @justusschock