[FlexAttention] Add mechanism to get optimal autotune decision#165817
Closed
drisspg wants to merge 7 commits intogh/drisspg/212/basefrom
Closed
[FlexAttention] Add mechanism to get optimal autotune decision#165817drisspg wants to merge 7 commits intogh/drisspg/212/basefrom
drisspg wants to merge 7 commits intogh/drisspg/212/basefrom
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/165817
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit c3fc9c4 with merge base 507614b ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Skylion007
reviewed
Oct 18, 2025
torch/_inductor/select_algorithm.py
Outdated
| return mm_file_name | ||
|
|
||
|
|
||
| _FLEX_ATTENTION_TUNABLE_KEYS = frozenset( |
Collaborator
There was a problem hiding this comment.
Is the better abstraction here just to create a typed_dict class and statically go over the keys maybe? With a default sentinel value? Or maybe that's just too complicated.
Skylion007
reviewed
Oct 18, 2025
Chillee
approved these changes
Oct 28, 2025
Contributor
Author
|
@pytorchbot merge |
Collaborator
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Khanaksahu
pushed a commit
to Khanaksahu/pytorch-fork
that referenced
this pull request
Nov 17, 2025
ghstack-source-id: 66998d1 Pull-Request: pytorch/pytorch#165817
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
4 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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):
Script: meta-pytorch/attention-gym#169
Feels directionally okay but there is some bike shedding / this could be quite prone to collision of keys depending on mask mod and score mod changes and simple cache key.
Usecase: meta-pytorch/attention-gym#169
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @chenyang78 @kadeng @muchulee8 @amjames @chauhang @aakhundov @coconutruben @Chillee @yanboliang @BoyuanFeng