This repository was archived by the owner on Nov 1, 2024. It is now read-only.
Fix the bug that the last checkpoint may not be saved properly in some cases#122
Open
0x4f5da2 wants to merge 1 commit intofacebookresearch:mainfrom
Open
Fix the bug that the last checkpoint may not be saved properly in some cases#1220x4f5da2 wants to merge 1 commit intofacebookresearch:mainfrom
0x4f5da2 wants to merge 1 commit intofacebookresearch:mainfrom
Conversation
facebook-github-bot
left a comment
There was a problem hiding this comment.
@rajprateek has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
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 subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
If a non-master process exit before the master process enter
save_checkpoint, the training process couldn't end properly. Addingtorch.distributed.barrier()will fix it.