Add FuseQATConvBN to fuse_ops (#19442)#19442
Merged
Merged
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/19442
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ❌ 1 New Failure, 5 Unrelated FailuresAs of commit ee9d4df with merge base fe98297 ( NEW FAILURE - The following job has failed:
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
|
@ethansfng has exported this pull request. If you are a Meta employee, you can view the originating Diff in D104497938. |
This PR needs a
|
meta-codesync
Bot
changed the title
ethansfng
added a commit
to ethansfng/executorch
that referenced
this pull request
May 10, 2026
Summary:
Adds a FuseQATConvBN which folds the QAT Conv-BN simulation chain (`conv → q → dq → div(scale) → add(orig_bias) → batch_norm`) inserted by `prepare_qat_pt2e` into the conv's quantized bias and removes the chain.
The pass runs in two steps inside a single `call()`:
1. Bias prep — for each conv, create a zero-filled quantized bias if missing, or quantize a float bias as per-tensor int32. Required so step 2 has a quantized bias slot to write the BN correction into.
2. Fold — for each matched chain, compute the BN correction
C = (orig_bias - running_mean) * bn_weight / sqrt(running_var + eps) + bn_bias
and absorb it into the conv's quantized bias in place. Erase the chain + batch_norm.
Differential Revision: D104497938
ethansfng
force-pushed
the
export-D104497938
branch
from
d5c07d4 to
84d7498
Compare
ethansfng
added a commit
to ethansfng/executorch
that referenced
this pull request
May 10, 2026
Summary:
Adds a FuseQATConvBN which folds the QAT Conv-BN simulation chain (`conv → q → dq → div(scale) → add(orig_bias) → batch_norm`) inserted by `prepare_qat_pt2e` into the conv's quantized bias and removes the chain.
The pass runs in two steps inside a single `call()`:
1. Bias prep — for each conv, create a zero-filled quantized bias if missing, or quantize a float bias as per-tensor int32. Required so step 2 has a quantized bias slot to write the BN correction into.
2. Fold — for each matched chain, compute the BN correction
C = (orig_bias - running_mean) * bn_weight / sqrt(running_var + eps) + bn_bias
and absorb it into the conv's quantized bias in place. Erase the chain + batch_norm.
Differential Revision: D104497938
ethansfng
force-pushed
the
export-D104497938
branch
from
84d7498 to
472a5cd
Compare
ethansfng
added a commit
to ethansfng/executorch
that referenced
this pull request
May 10, 2026
Summary:
Adds a FuseQATConvBN which folds the QAT Conv-BN simulation chain (`conv → q → dq → div(scale) → add(orig_bias) → batch_norm`) inserted by `prepare_qat_pt2e` into the conv's quantized bias and removes the chain.
The pass runs in two steps inside a single `call()`:
1. Bias prep — for each conv, create a zero-filled quantized bias if missing, or quantize a float bias as per-tensor int32. Required so step 2 has a quantized bias slot to write the BN correction into.
2. Fold — for each matched chain, compute the BN correction
C = (orig_bias - running_mean) * bn_weight / sqrt(running_var + eps) + bn_bias
and absorb it into the conv's quantized bias in place. Erase the chain + batch_norm.
Differential Revision: D104497938
ethansfng
force-pushed
the
export-D104497938
branch
2 times, most recently
from
c5e9566 to
0ced839
Compare
ethansfng
added a commit
to ethansfng/executorch
that referenced
this pull request
May 11, 2026
Summary:
Adds a FuseQATConvBN which folds the QAT Conv-BN simulation chain (`conv → q → dq → div(scale) → add(orig_bias) → batch_norm`) inserted by `prepare_qat_pt2e` into the conv's quantized bias and removes the chain.
The pass runs in two steps inside a single `call()`:
1. Bias prep — for each conv, create a zero-filled quantized bias if missing, or quantize a float bias as per-tensor int32. Required so step 2 has a quantized bias slot to write the BN correction into.
2. Fold — for each matched chain, compute the BN correction
C = (orig_bias - running_mean) * bn_weight / sqrt(running_var + eps) + bn_bias
and absorb it into the conv's quantized bias in place. Erase the chain + batch_norm.
Differential Revision: D104497938
ethansfng
force-pushed
the
export-D104497938
branch
from
0ced839 to
51cb526
Compare
ethansfng
added a commit
to ethansfng/executorch
that referenced
this pull request
May 12, 2026
Summary:
Adds a FuseQATConvBN which folds the QAT Conv-BN simulation chain (`conv → q → dq → div(scale) → add(orig_bias) → batch_norm`) inserted by `prepare_qat_pt2e` into the conv's quantized bias and removes the chain.
The pass runs in two steps inside a single `call()`:
1. Bias prep — for each conv, create a zero-filled quantized bias if missing, or quantize a float bias as per-tensor int32. Required so step 2 has a quantized bias slot to write the BN correction into.
2. Fold — for each matched chain, compute the BN correction
C = (orig_bias - running_mean) * bn_weight / sqrt(running_var + eps) + bn_bias
and absorb it into the conv's quantized bias in place. Erase the chain + batch_norm.
Differential Revision: D104497938
ethansfng
added a commit
to ethansfng/executorch
that referenced
this pull request
May 12, 2026
Summary:
Adds a FuseQATConvBN which folds the QAT Conv-BN simulation chain (`conv → q → dq → div(scale) → add(orig_bias) → batch_norm`) inserted by `prepare_qat_pt2e` into the conv's quantized bias and removes the chain.
The pass runs in two steps inside a single `call()`:
1. Bias prep — for each conv, create a zero-filled quantized bias if missing, or quantize a float bias as per-tensor int32. Required so step 2 has a quantized bias slot to write the BN correction into.
2. Fold — for each matched chain, compute the BN correction
C = (orig_bias - running_mean) * bn_weight / sqrt(running_var + eps) + bn_bias
and absorb it into the conv's quantized bias in place. Erase the chain + batch_norm.
Reviewed By: DrJessop
Differential Revision: D104497938
ethansfng
force-pushed
the
export-D104497938
branch
from
51cb526 to
b08ceb7
Compare
aliafzal
approved these changes
May 12, 2026
Summary:
Adds a FuseQATConvBN which folds the QAT Conv-BN simulation chain (`conv → q → dq → div(scale) → add(orig_bias) → batch_norm`) inserted by `prepare_qat_pt2e` into the conv's quantized bias and removes the chain.
The pass runs in two steps inside a single `call()`:
1. Bias prep — for each conv, create a zero-filled quantized bias if missing, or quantize a float bias as per-tensor int32. Required so step 2 has a quantized bias slot to write the BN correction into.
2. Fold — for each matched chain, compute the BN correction
C = (orig_bias - running_mean) * bn_weight / sqrt(running_var + eps) + bn_bias
and absorb it into the conv's quantized bias in place. Erase the chain + batch_norm.
Reviewed By: DrJessop
Differential Revision: D104497938
ethansfng
force-pushed
the
export-D104497938
branch
from
b08ceb7 to
ee9d4df
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:
Adds a FuseQATConvBN which folds the QAT Conv-BN simulation chain (
conv → q → dq → div(scale) → add(orig_bias) → batch_norm) inserted byprepare_qat_pt2einto the conv's quantized bias and removes the chain.The pass runs in two steps inside a single
call():C = (orig_bias - running_mean) * bn_weight / sqrt(running_var + eps) + bn_bias
and absorb it into the conv's quantized bias in place. Erase the chain + batch_norm.
Reviewed By: DrJessop
Differential Revision: D104497938