Sync: rank-0 CIFAR-10 download + barrier (fix examples gloo timeout)#270
Open
wz337 wants to merge 1 commit into
Open
Sync: rank-0 CIFAR-10 download + barrier (fix examples gloo timeout)#270wz337 wants to merge 1 commit into
wz337 wants to merge 1 commit into
Conversation
Summary: Regenerated OSS tree (oss_sync.sh) to propagate the dev-source fix in the parent diff: the CIFAR-10 example now downloads on rank 0 only and barriers before all ranks load the shared dataset with download=False. This fixes the persistent `examples / Python 3.12` GitHub CI failure where concurrent per-rank downloads throttled + desynced the gloo ranks, timing out the all-reduce in loss.backward() after the 30-min deadline. Only examples/utils.py changes in content; the rest of the tree is unchanged. Differential Revision: D112649967
meta-cla
Bot
added
the
CLA Signed
|
@wz337 has exported this pull request. If you are a Meta employee, you can view the originating Diff in D112649967. |
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary:
Regenerated OSS tree (oss_sync.sh) to propagate the dev-source fix in the
parent diff: the CIFAR-10 example now downloads on rank 0 only and barriers
before all ranks load the shared dataset with download=False. This fixes the
persistent
examples / Python 3.12GitHub CI failure where concurrentper-rank downloads throttled + desynced the gloo ranks, timing out the
all-reduce in loss.backward() after the 30-min deadline.
Only examples/utils.py changes in content; the rest of the tree is unchanged.
Differential Revision: D112649967