Avoid device sync in training loss accumulation#44123
Open
cyyever wants to merge 3 commits intohuggingface:mainfrom
Open
Avoid device sync in training loss accumulation#44123cyyever wants to merge 3 commits intohuggingface:mainfrom
cyyever wants to merge 3 commits intohuggingface:mainfrom
Conversation
Member
|
Hey @cyyever! We did some internal discussion about CUDA sync PRs - would you be able to make a benchmark or profile graph showing the reduction in syncs? We might make those mandatory in future just because it can sometimes be hard to tell from reading the code where exactly syncs would occur |
Signed-off-by: Yuanyuan Chen <cyyever@outlook.com>
Signed-off-by: Yuanyuan Chen <cyyever@outlook.com>
Signed-off-by: Yuanyuan Chen <cyyever@outlook.com>
This was referenced Apr 29, 2026
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What does this PR do?
This PR avoids device sync in training loss accumulation by
torch.where. Theis_torch_xla_availablecondition is also removed.