fix transpose / permutations fusion pass #10780
Merged
Conversation
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
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/10780
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 684edfc with merge base b1b46ee ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D74447383 |
mcremon-meta
approved these changes
May 9, 2025
Summary: Follow up on D73619452 We noticed that this pattern used to be fused, but stopped being fused after D73619452. What happens is that we were checking for unitary dimensions. So a tensor of shape [1, 5, 30] is equivalent (in memory) to [5, 1, 30]. So, transposing on [0, 2] then on [0, 1] is (almost) equivalent to an identity and can be fused This diff re-establishes this behavior. Reviewed By: mcremon-meta Differential Revision: D74447383
ThomasJannaud
force-pushed
the
export-D74447383
branch
from
5af192d to
684edfc
Compare
Contributor
|
This pull request was exported from Phabricator. Differential Revision: D74447383 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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Outdated suggestions cannot be applied.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.
Summary:
Follow up on D73619452
We noticed that this pattern used to be fused, but stopped being fused after D73619452.
What happens is that we were checking for unitary dimensions. So a tensor of shape [1, 5, 30] is equivalent (in memory) to [5, 1, 30]. So, transposing on [0, 2] then on [0, 1] is (almost) equivalent to an identity and can be fused
This diff re-establishes this behavior.
Differential Revision: D74447383