Lower Conv1d atomically through TOSA Conv2d (#22282) - #22282
Open
mcremon-meta wants to merge 1 commit into
Open
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/22282
Note: Links to docs will display an error until the docs builds have been completed. ⏳ 190 Pending, 2 Unrelated FailuresAs of commit 8499c66 with merge base c27baa8 ( BROKEN TRUNK - The following jobs failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
Bot
added
the
CLA Signed
Contributor
|
@mcremon-meta has exported this pull request. If you are a Meta employee, you can view the originating Diff in D116686805. |
This PR needs a
|
Contributor
meta-codesync
Bot
changed the title
meta-codesync Bot
pushed a commit
that referenced
this pull request
Aug 28, 2026
Summary: Lower one-dimensional convolutions atomically during convolution rewriting instead of expanding them to rank four in an earlier pass. Emit rank-three layout boundaries around the operation and introduce the singleton spatial dimension only after converting the input to the backend layout. Rewrite regular and depthwise weights directly into their TOSA Conv2d layouts, and retain the earlier expansion path only for transposed one-dimensional convolutions. This exposes rank-three transforms to the existing layout propagation and cancellation passes while keeping the generated convolution representation valid. Differential Revision: D116686805
meta-codesync
Bot
force-pushed
the
export-D116686805
branch
from
August 28, 2026 23:44
9d1d34a to
dd3785c
Compare
meta-codesync Bot
pushed a commit
that referenced
this pull request
Aug 29, 2026
Summary: Lower one-dimensional convolutions atomically during convolution rewriting instead of expanding them to rank four in an earlier pass. Emit rank-three layout boundaries around the operation and introduce the singleton spatial dimension only after converting the input to the backend layout. Rewrite regular and depthwise weights directly into their TOSA Conv2d layouts, and retain the earlier expansion path only for transposed one-dimensional convolutions. This exposes rank-three transforms to the existing layout propagation and cancellation passes while keeping the generated convolution representation valid. Differential Revision: D116686805
meta-codesync
Bot
force-pushed
the
export-D116686805
branch
from
August 29, 2026 04:01
dd3785c to
8fe0e49
Compare
meta-codesync Bot
pushed a commit
that referenced
this pull request
Aug 29, 2026
Summary: Lower one-dimensional convolutions atomically during convolution rewriting instead of expanding them to rank four in an earlier pass. Emit rank-three layout boundaries around the operation and introduce the singleton spatial dimension only after converting the input to the backend layout. Rewrite regular and depthwise weights directly into their TOSA Conv2d layouts, and retain the earlier expansion path only for transposed one-dimensional convolutions. This exposes rank-three transforms to the existing layout propagation and cancellation passes while keeping the generated convolution representation valid. Differential Revision: D116686805
meta-codesync
Bot
force-pushed
the
export-D116686805
branch
from
August 29, 2026 16:31
8fe0e49 to
40a5b10
Compare
meta-codesync Bot
pushed a commit
that referenced
this pull request
Aug 30, 2026
Summary: Lower one-dimensional convolutions atomically during convolution rewriting instead of expanding them to rank four in an earlier pass. Emit rank-three layout boundaries around the operation and introduce the singleton spatial dimension only after converting the input to the backend layout. Rewrite regular and depthwise weights directly into their TOSA Conv2d layouts, and retain the earlier expansion path only for transposed one-dimensional convolutions. This exposes rank-three transforms to the existing layout propagation and cancellation passes while keeping the generated convolution representation valid. Differential Revision: D116686805
meta-codesync
Bot
force-pushed
the
export-D116686805
branch
from
August 30, 2026 01:00
40a5b10 to
b033451
Compare
meta-codesync Bot
pushed a commit
that referenced
this pull request
Aug 30, 2026
Summary: Lower one-dimensional convolutions atomically during convolution rewriting instead of expanding them to rank four in an earlier pass. Emit rank-three layout boundaries around the operation and introduce the singleton spatial dimension only after converting the input to the backend layout. Rewrite regular and depthwise weights directly into their TOSA Conv2d layouts, and retain the earlier expansion path only for transposed one-dimensional convolutions. This exposes rank-three transforms to the existing layout propagation and cancellation passes while keeping the generated convolution representation valid. Differential Revision: D116686805
meta-codesync
Bot
force-pushed
the
export-D116686805
branch
from
August 30, 2026 05:16
b033451 to
6326254
Compare
Summary: Lower one-dimensional convolutions atomically during convolution rewriting instead of expanding them to rank four in an earlier pass. Emit rank-three layout boundaries around the operation and introduce the singleton spatial dimension only after converting the input to the backend layout. Rewrite regular and depthwise weights directly into their TOSA Conv2d layouts, and retain the earlier expansion path only for transposed one-dimensional convolutions. This exposes rank-three transforms to the existing layout propagation and cancellation passes while keeping the generated convolution representation valid. Differential Revision: D116686805
meta-codesync Bot
pushed a commit
that referenced
this pull request
Aug 30, 2026
Summary: Lower one-dimensional convolutions atomically during convolution rewriting instead of expanding them to rank four in an earlier pass. Emit rank-three layout boundaries around the operation and introduce the singleton spatial dimension only after converting the input to the backend layout. Rewrite regular and depthwise weights directly into their TOSA Conv2d layouts, and retain the earlier expansion path only for transposed one-dimensional convolutions. This exposes rank-three transforms to the existing layout propagation and cancellation passes while keeping the generated convolution representation valid. Differential Revision: D116686805
meta-codesync
Bot
force-pushed
the
export-D116686805
branch
from
August 30, 2026 19:26
6326254 to
9fcf849
Compare
meta-codesync
Bot
force-pushed
the
export-D116686805
branch
from
August 30, 2026 19:26
9fcf849 to
8499c66
Compare
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
2 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:
Lower one-dimensional convolutions atomically during convolution rewriting instead of expanding them to rank four in an earlier pass. Emit rank-three layout boundaries around the operation and introduce the singleton spatial dimension only after converting the input to the backend layout.
Rewrite regular and depthwise weights directly into their TOSA Conv2d layouts, and retain the earlier expansion path only for transposed one-dimensional convolutions. This exposes rank-three transforms to the existing layout propagation and cancellation passes while keeping the generated convolution representation valid.
Differential Revision: D116686805