Fix misplaced overflow handling return in fused_optimizer.py #7645
Conversation
This file contains hidden or 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
Contributor
|
Thanks! This should fix #7632. |
Contributor
Author
|
Hi @tjruwase, could you please help in reviewing this PR? |
tjruwase
approved these changes
Oct 28, 2025
Contributor
Author
|
Hi @tohtana, could you please help in reviewing this PR? |
rraminen
added a commit
to rraminen/DeepSpeed
that referenced
this pull request
Dec 1, 2025
…edai#7645) This PR fixes an issue in deepspeed/runtime/fp16/fused_optimizer.py where the gradient overflow handling logic incorrectly exited the function too early, resulting in wrong forward pass and loss calculations in certain FP16 training scenarios. The `return self.overflow` and `self.timers.log(OVERFLOW_TIMERS)` calls are now correctly moved inside the `if self.overflow:` block so that the function only returns early when an actual overflow is detected. Origin of the error: deepspeedai@889f0ea cc: @jithunnair-amd Co-authored-by: Olatunji Ruwase <[email protected]> Signed-off-by: rraminen <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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Outdated suggestions cannot be applied.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.
This PR fixes an issue in deepspeed/runtime/fp16/fused_optimizer.py where the gradient overflow handling logic incorrectly exited the function too early, resulting in wrong forward pass and loss calculations in certain FP16 training scenarios.
The
return self.overflowandself.timers.log(OVERFLOW_TIMERS)calls are now correctly moved inside theif self.overflow:block so that the function only returns early when an actual overflow is detected.Origin of the error: 889f0ea
cc: @jithunnair-amd