This repository has been archived by the owner on Jul 1, 2024. It is now read-only.
Support LARC for SGD optimizer only in classy vision #408
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: D20139718 |
2 similar comments
|
This pull request was exported from Phabricator. Differential Revision: D20139718 |
|
This pull request was exported from Phabricator. Differential Revision: D20139718 |
…h#408) Summary: Pull Request resolved: facebookresearch#408 Pull Request resolved: fairinternal/ClassyVision#64 In an attempt to implement SimpleCLR for contrastive losses, I needed LARC to enable large batch training. mannatsingh had already done work on this during classy vision open source release. https://our.intern.facebook.com/intern/diff/D18542126/ I initially tried using that diff to have a separate standalone LARC to work for any optimizer but it turned out to be tricky to setup correctly as we need to wrap a given optimizer in LARC (the `getattr` and `setattr` functions were not working). I talked to vreis about it and we decided that for now, we can support it for SGD only, file a task to support other optimizers too later after discussions with mannatsingh once he's back. Differential Revision: D20139718 fbshipit-source-id: c8cf4d545e6ce94cca8e646f68d519197856f675
prigoyal
force-pushed
the
export-D20139718
branch
from
6d2f06c
to
4add1b3
Compare
|
This pull request was exported from Phabricator. Differential Revision: D20139718 |
|
This pull request has been merged in 04a99c8. |
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Outdated suggestions cannot be applied.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.
Summary:
In an attempt to implement SimpleCLR for contrastive losses, I needed LARC to enable large batch training. mannatsingh had already done work on this during classy vision open source release. https://our.intern.facebook.com/intern/diff/D18542126/
I initially tried using that diff to have a separate standalone LARC to work for any optimizer but it turned out to be tricky to setup correctly as we need to wrap a given optimizer in LARC (the
getattrandsetattrfunctions were not working). I talked to vreis about it and we decided that for now, we can support it for SGD only, file a task to support other optimizers too later after discussions with mannatsingh once he's back.Differential Revision: D20139718