Add 16A8W linear ops support and test by Ninja91 · Pull Request #13658 · pytorch/executorch

Ninja91 · 2025-08-25T21:41:01Z

Stack from ghstack (oldest at bottom):

Adds linear ops test using the 16A8W config in INT16 profile.
Adds support in view ops validation for INT16 Dtype.
Validated with TOSA pipeline test.
Checked earlier marked flaky tests no longer flaky and remove markers.

Note: Not verified with tosa reference model run.

Differential Revision: D80308822

- Adds linear ops test using the 16A8W config in INT16 profile. - Adds support in view ops validation for INT16 Dtype. - Validated with TOSA pipeline test. - Checked earlier marked flaky tests no longer flaky and remove markers. Note: Not verified with tosa reference model run. Differential Revision: [D80308822](https://our.internmc.facebook.com/intern/diff/D80308822/) [ghstack-poisoned]

pytorch-bot · 2025-08-25T21:41:04Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/13658

📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 3 New Failures

As of commit 019d222 with merge base 9053089 ():

NEW FAILURES - The following jobs have failed:

Build documentation / build (buck2) / Build doc (gh)
At least one of the pre-conditions you specified did not hold
pull / unittest-arm-backend-with-no-fvp (test_pytest_models) / linux-job (gh)
backends/arm/test/models/stable_diffusion/test_T5EncoderModel.py::TestT5EncoderModel::test_T5EncoderModel_tosa_MI
pull / unittest-arm-backend-with-no-fvp (test_pytest_ops) / linux-job (gh)
RuntimeError: Command docker exec -t 1ad9c186779ac31fa9b43e1c5bc7dcec8073d0a91871ee882f9acf9a1446781e /exec failed with exit code 1

This comment was automatically generated by Dr. CI and updates every 15 minutes.

- Adds linear ops test using the 16A8W config in INT16 profile. - Adds support in view ops validation for INT16 Dtype. - Validated with TOSA pipeline test. - Checked earlier marked flaky tests no longer flaky and remove markers. Note: Not verified with tosa reference model run. Differential Revision: [D80308822](https://our.internmc.facebook.com/intern/diff/D80308822/) ghstack-source-id: 305494502 Pull Request resolved: #13658

facebook-github-bot · 2025-08-25T21:41:26Z

This pull request was exported from Phabricator. Differential Revision: D80308822

github-actions · 2025-08-25T21:41:57Z

This PR needs a `release notes:` label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

Ninja91 · 2025-08-25T21:43:47Z

Updated version of #13448. Closing the former.

Ninja91 · 2025-08-25T22:18:54Z

@per @digantdesai some of CI is failing but it doesn't appear to be related to code change here.

- Adds linear ops test using the 16A8W config in INT16 profile. - Adds support in view ops validation for INT16 Dtype. - Validated with TOSA pipeline test. - Checked earlier marked flaky tests no longer flaky and remove markers. Note: Not verified with tosa reference model run. Differential Revision: [D80308822](https://our.internmc.facebook.com/intern/diff/D80308822/) [ghstack-poisoned]

facebook-github-bot · 2025-08-26T04:16:34Z

This pull request was exported from Phabricator. Differential Revision: D80308822

per

Nice! Two small suggestions, otherwise LGTM.

backends/arm/operators/op_transpose.py

per · 2025-08-26T07:29:47Z

backends/arm/operators/op_view.py

        validate_valid_dtype(
            self.target,
            [inputs[0], output],
-            [ts.DType.INT8, ts.DType.INT32, ts.DType.FP32, ts.DType.BOOL],


Same as for op_transpose.py

backends/arm/test/ops/test_linear.py

digantdesai · 2025-08-26T15:20:11Z

backends/arm/test/ops/test_sigmoid_16bit.py::test_sigmoid_tosa_INT_add_sigmoid[rand] failure seems unrelated.

- Adds linear ops test using the 16A8W config in INT16 profile. - Adds support in view ops validation for INT16 Dtype. - Validated with TOSA pipeline test. - Checked earlier marked flaky tests no longer flaky and remove markers. Note: Not verified with tosa reference model run. Differential Revision: [D80308822](https://our.internmc.facebook.com/intern/diff/D80308822/) [ghstack-poisoned]

facebook-github-bot · 2025-08-26T23:37:26Z

This pull request was exported from Phabricator. Differential Revision: D80308822

- Adds linear ops test using the 16A8W config in INT16 profile. - Adds support in view ops validation for INT16 Dtype. - Validated with TOSA pipeline test. - Checked earlier marked flaky tests no longer flaky and remove markers. Note: Not verified with tosa reference model run. Differential Revision: [D80308822](https://our.internmc.facebook.com/intern/diff/D80308822/) [ghstack-poisoned]

Pull Request resolved: #13658 - Adds linear ops test using the 16A8W config in INT16 profile. - Adds support in view ops validation for INT16 Dtype. - Validated with TOSA pipeline test. - Checked earlier marked flaky tests no longer flaky and remove markers. Note: Not verified with tosa reference model run. ghstack-source-id: 305893128 Differential Revision: [D80308822](https://our.internmc.facebook.com/intern/diff/D80308822/)

facebook-github-bot · 2025-08-27T06:26:09Z

This pull request was exported from Phabricator. Differential Revision: D80308822

- Adds linear ops test using the 16A8W config in INT16 profile. - Adds support in view ops validation for INT16 Dtype. - Validated with TOSA pipeline test. - Checked earlier marked flaky tests no longer flaky and remove markers. Note: Not verified with tosa reference model run. Differential Revision: [D80308822](https://our.internmc.facebook.com/intern/diff/D80308822/) [ghstack-poisoned]

facebook-github-bot · 2025-08-27T06:30:57Z

This pull request was exported from Phabricator. Differential Revision: D80308822

per · 2025-08-29T07:20:57Z

@digantdesai Were was this merged? Can't see it on main... Getting a bit confused by the ghstack stuff 😕

digantdesai · 2025-08-29T15:30:22Z

@Ninja91 ?

Ninja91 · 2025-09-01T07:25:56Z

@digantdesai Were was this merged? Can't see it on main... Getting a bit confused by the ghstack stuff 😕

This diff was merged and new PR was created for main branch here: #13754

I see that there are some tests failing. Looks like the change to use support_integer() causes FP tests fail as some ops are still INT.
if self.tosa_spec.support_integer(): valid.dtypes.extend([ts.DType.INT8, ts.DType.INT16]) if self.tosa_spec.support_float(): valid.dtypes.extend([ts.DType.FP16, ts.DType.FP32])

What would be your suggestion?

digantdesai · 2025-09-02T21:09:45Z

this didn't merge with main, can you push this to main manually?

@Ninja91

This PR was created by the merge bot to help merge the original PR into the main branch. ghstack PR number: #13658 by @Ninja91 ^ Please use this as the source of truth for the PR details, comments, and reviews ghstack PR base: https://github.com/pytorch/executorch/tree/gh/Ninja91/3/base ghstack PR head: https://github.com/pytorch/executorch/tree/gh/Ninja91/3/head Merge bot PR base: https://github.com/pytorch/executorch/tree/gh/Ninja91/1/orig Merge bot PR head: https://github.com/pytorch/executorch/tree/gh/Ninja91/3/orig @diff-train-skip-merge cc @digantdesai @freddan80 @per @zingo @oscarandersson8218 --------- Co-authored-by: Nitin Jain <jainnitin@meta.com>

Ninja91 requested a review from digantdesai as a code owner August 25, 2025 21:41

Ninja91 mentioned this pull request Aug 25, 2025

Add 16A8W quantization configuration utility for ARM backend #13641

Merged

meta-cla bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Aug 25, 2025

facebook-github-bot added the fb-exported label Aug 25, 2025

Ninja91 self-assigned this Aug 25, 2025

Ninja91 requested a review from per August 25, 2025 21:43

This was referenced Aug 25, 2025

Add 16A8W linear ops support and test #13448

Closed

Add 16A8W linear ops support and test #13635

Closed

Ninja91 mentioned this pull request Aug 25, 2025

Add 16A8W support and test for add operation #13653

Closed

per approved these changes Aug 26, 2025

View reviewed changes

digantdesai reviewed Aug 26, 2025

View reviewed changes

backends/arm/test/ops/test_linear.py Show resolved Hide resolved

facebook-github-bot merged commit 8d7c687 into gh/Ninja91/3/base Aug 28, 2025
107 of 112 checks passed

facebook-github-bot deleted the gh/Ninja91/3/head branch August 28, 2025 07:27

facebook-github-bot temporarily deployed to cherry-pick-bot August 28, 2025 07:27 — with GitHub Actions Inactive

pytorchbot mentioned this pull request Aug 28, 2025

Arm backend: Add 16A8W linear ops support and test #13754

Merged

Ninja91 mentioned this pull request Sep 2, 2025

Arm backend: Add passes to handle int64 const and int64 output ops #13803

Merged

Conversation

Ninja91 commented Aug 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Aug 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/13658

❌ 3 New Failures

Uh oh!

facebook-github-bot commented Aug 25, 2025

Uh oh!

github-actions bot commented Aug 25, 2025

This PR needs a release notes: label

Uh oh!

Ninja91 commented Aug 25, 2025

Uh oh!

Ninja91 commented Aug 25, 2025

Uh oh!

facebook-github-bot commented Aug 26, 2025

Uh oh!

per left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

per Aug 26, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

digantdesai commented Aug 26, 2025

Uh oh!

facebook-github-bot commented Aug 26, 2025

Uh oh!

facebook-github-bot commented Aug 27, 2025

Uh oh!

facebook-github-bot commented Aug 27, 2025

Uh oh!

Uh oh!

per commented Aug 29, 2025

Uh oh!

digantdesai commented Aug 29, 2025

Uh oh!

Ninja91 commented Sep 1, 2025

Uh oh!

digantdesai commented Sep 2, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Ninja91 commented Aug 25, 2025 •

edited

Loading

pytorch-bot bot commented Aug 25, 2025 •

edited

Loading

This PR needs a `release notes:` label