Refactor quantizer: Only replace with per-tensor variants #14974
+179
−485
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/14974
Note: Links to docs will display an error until the docs builds have been completed. This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
bot
added
the
CLA Signed
This PR needs a
|
DrJessop
force-pushed
the
export-D83873738
branch
from
c434d1e to
3e65bea
Compare
DrJessop
pushed a commit
to DrJessop/executorch
that referenced
this pull request
Oct 10, 2025
) Summary: In our previous flow, we would replace ops with default variants, have a special fusion pass which constructs singleton tensors for a variety of fused quantized ops, and then we would call a replace ops to turn them into per-tensor-variants. I confirmed this was for legacy reasons, so a cleanup was much due. This diff also fixes any ref implementations during the refactor. Reviewed By: zonglinpeng Differential Revision: D83873738
DrJessop
force-pushed
the
export-D83873738
branch
from
3e65bea to
33a47e1
Compare
DrJessop
pushed a commit
to DrJessop/executorch
that referenced
this pull request
Oct 10, 2025
) Summary: In our previous flow, we would replace ops with default variants, have a special fusion pass which constructs singleton tensors for a variety of fused quantized ops, and then we would call a replace ops to turn them into per-tensor-variants. I confirmed this was for legacy reasons, so a cleanup was much due. This diff also fixes any ref implementations during the refactor. Reviewed By: zonglinpeng Differential Revision: D83873738
DrJessop
force-pushed
the
export-D83873738
branch
from
33a47e1 to
43e31be
Compare
DrJessop
pushed a commit
to DrJessop/executorch
that referenced
this pull request
Oct 10, 2025
) Summary: In our previous flow, we would replace ops with default variants, have a special fusion pass which constructs singleton tensors for a variety of fused quantized ops, and then we would call a replace ops to turn them into per-tensor-variants. I confirmed this was for legacy reasons, so a cleanup was much due. This diff also fixes any ref implementations during the refactor. Reviewed By: zonglinpeng Differential Revision: D83873738
DrJessop
pushed a commit
to DrJessop/executorch
that referenced
this pull request
Oct 10, 2025
) Summary: In our previous flow, we would replace ops with default variants, have a special fusion pass which constructs singleton tensors for a variety of fused quantized ops, and then we would call a replace ops to turn them into per-tensor-variants. I confirmed this was for legacy reasons, so a cleanup was much due. This diff also fixes any ref implementations during the refactor. Reviewed By: zonglinpeng Differential Revision: D83873738
DrJessop
force-pushed
the
export-D83873738
branch
2 times, most recently
from
f84f831 to
271538a
Compare
DrJessop
pushed a commit
to DrJessop/executorch
that referenced
this pull request
Oct 10, 2025
) Summary: In our previous flow, we would replace ops with default variants, have a special fusion pass which constructs singleton tensors for a variety of fused quantized ops, and then we would call a replace ops to turn them into per-tensor-variants. I confirmed this was for legacy reasons, so a cleanup was much due. This diff directly replaces ops with the per-tensor variants and removes the pass which replaces singleton tensors with scalars. Reviewed By: zonglinpeng Differential Revision: D83873738
added 2 commits
October 10, 2025 15:09
Summary: Matmul was relying on linear infra which didn't support batched second argument. This adds support. Differential Revision: D84279595
) Summary: In our previous flow, we would replace ops with default variants, have a special fusion pass which constructs singleton tensors for a variety of fused quantized ops, and then we would call a replace ops to turn them into per-tensor-variants. I confirmed this was for legacy reasons, so a cleanup was much due. This diff directly replaces ops with the per-tensor variants and removes the pass which replaces singleton tensors with scalars. Reviewed By: zonglinpeng Differential Revision: D83873738
DrJessop
force-pushed
the
export-D83873738
branch
from
271538a to
c805660
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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 our previous flow, we would replace ops with default variants, have a special fusion pass which constructs singleton tensors for a variety of fused quantized ops, and then we would call a replace ops to turn them into per-tensor-variants.
I confirmed this was for legacy reasons, so a cleanup was much due.
This diff also fixes any ref implementations during the refactor.
Reviewed By: zonglinpeng
Differential Revision: D83873738