-
Notifications
You must be signed in to change notification settings - Fork 25.7k
[FSDP2] Fixed mistargeted backward prefetch #137348
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
Conversation
[ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/137348
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 4fb18f1 with merge base d1b87e2 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
cc XilunWu H-Huang kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o [ghstack-poisoned]
cc XilunWu H-Huang kwen2501 wanchaol fegin fduwjj wz337 wconstab d4l3k c-p-i-o [ghstack-poisoned]
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nice catch!
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
ghstack-source-id: 60f148d Pull Request resolved: pytorch/pytorch#137348
Stack from ghstack (oldest at bottom):
If there is an
unshard(top-half) without await_for_unshard(bottom-half), then the next iteration'sunshardwill be a no-op. This can unexpectedly not propagate the optimizer update on the sharded parameters to the unsharded parameters, so it is better to clear thatunshardat the end of backward.cc @XilunWu @H-Huang @kwen2501 @wanchaol @fegin @fduwjj @wz337 @wconstab @d4l3k @c-p-i-o