Arm backend: Avoid shared qspec bridges for uint8 IO#20976
Merged
bdemirb merged 4 commits intoJul 17, 2026
Conversation
When preserved uint8 model IO is used, SharedQspecQuantizer can merge image-scale tensors with wider internal activation ranges. Avoid using cat, concatenate, stack, pixel_shuffle, and slice as fallback shared-qspec bridge ops for graphs with uint8 IO. This keeps normal int8 quantization behaviour unchanged while preserving image input qparams for uint8 IO flows. Add a regression test covering image-like uint8 IO joined with a high-range branch. Signed-off-by: Baris Demir <baris.demir@arm.com> Change-Id: Ieeb1cf59e606674ac34050ebe21fc080740cc567
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/20976
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit be2e87b with merge base f6cef83 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
Bot
added
the
CLA Signed
Collaborator
Author
|
@pytorchbot label "partner: arm" |
pytorch-bot
Bot
added
the
partner: arm
Collaborator
Author
|
@pytorchbot label "release notes: arm" |
robell
approved these changes
Jul 17, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
When preserved uint8 model IO is used, SharedQspecQuantizer can merge image-scale tensors with wider internal activation ranges.
Avoid using cat, concatenate, stack, pixel_shuffle, and slice as fallback shared-qspec bridge ops for graphs with uint8 IO. This keeps normal int8 quantisation behaviour unchanged while preserving image input qparams for uint8 IO flows.
Add a regression test covering image-like uint8 IO joined with a high-range branch.
cc @digantdesai @freddan80 @per @zingo @oscarandersson8218 @mansnils @Sebastian-Larsson @robell @rascani