Use enable_gqa in place of repeat_kv
#641
Draft
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
[ghstack-poisoned]
awgu
added a commit
that referenced
this pull request
Oct 22, 2024
ghstack-source-id: e8781d3e797737c073bc487197ce804bce15502c Pull Request resolved: #641
facebook-github-bot
added
the
CLA Signed
|
cc., @XilunWu We may want to test if CP still works after this PR is landed. |
|
Lets wait before landing this, still working on the CuDNN backend integration for this feature and want to make sure everything is working accordingly |
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Outdated suggestions cannot be applied.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.
Stack from ghstack (oldest at bottom):
enable_gqain place ofrepeat_kv#641Saves some memory even when compile is enabled, MFU is roughly similar
pending some investigation from @drisspg
currently looks like cudnn attention with
enable_gqa=Truemay be recomputing therepeat_kvin backward, leading to memory savings, but no change in computation