Deepcopy FP module even if on meta device #1676
Closed
Conversation
This file contains 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
facebook-github-bot
added
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D53294788 |
s4ayub
added a commit
to s4ayub/torchrec-3
that referenced
this pull request
Feb 1, 2024
Summary: When we fx trace, even if there are 2 FP modules (because 2 cards), since it was sharded on meta, the ranks just have a reference to the FP on rank 0 and for whatever reason, FX eliminates the FP on rank 1 and it just shows the one on rank 0 do a deepcopy even when on meta device so each rank explicitly has their own copy, fx will persist it Reviewed By: lequytra, tissue3 Differential Revision: D53294788
s4ayub
force-pushed
the
export-D53294788
branch
from
77b3ff5 to
613556c
Compare
|
This pull request was exported from Phabricator. Differential Revision: D53294788 |
1 similar comment
|
This pull request was exported from Phabricator. Differential Revision: D53294788 |
s4ayub
added a commit
to s4ayub/torchrec-3
that referenced
this pull request
Feb 1, 2024
Summary: Pull Request resolved: pytorch#1676 When we fx trace, even if there are 2 FP modules (because 2 cards), since it was sharded on meta, the ranks just have a reference to the FP on rank 0 and for whatever reason, FX eliminates the FP on rank 1 and it just shows the one on rank 0 do a deepcopy even when on meta device so each rank explicitly has their own copy, fx will persist it Reviewed By: lequytra, tissue3 Differential Revision: D53294788 fbshipit-source-id: f70008a79e5f9fba2499d748d23587b1fc3c3c4a
s4ayub
force-pushed
the
export-D53294788
branch
from
613556c to
f27ca10
Compare
Summary: When we fx trace, even if there are 2 FP modules (because 2 cards), since it was sharded on meta, the ranks just have a reference to the FP on rank 0 and for whatever reason, FX eliminates the FP on rank 1 and it just shows the one on rank 0 do a deepcopy even when on meta device so each rank explicitly has their own copy, fx will persist it Reviewed By: lequytra, tissue3 Differential Revision: D53294788
s4ayub
force-pushed
the
export-D53294788
branch
from
f27ca10 to
15ba21c
Compare
|
This pull request was exported from Phabricator. Differential Revision: D53294788 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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:
When we fx trace, even if there are 2 FP modules (because 2 cards), since it was sharded on meta, the ranks just have a reference to the FP on rank 0
and for whatever reason, FX eliminates the FP on rank 1 and it just shows the one on rank 0
do a deepcopy even when on meta device so each rank explicitly has their own copy, fx will persist it
Reviewed By: lequytra, tissue3
Differential Revision: D53294788