Skip to content

Conversation

Ninja91
Copy link
Contributor

@Ninja91 Ninja91 commented Aug 15, 2025

Summary:

  • 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.

Note: Not verified with tosa reference model run.

Differential Revision: D80308822

cc @digantdesai @freddan80 @per @zingo @oscarandersson8218

@Ninja91 Ninja91 requested a review from digantdesai as a code owner August 15, 2025 03:15
Copy link

pytorch-bot bot commented Aug 15, 2025

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/13448

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

❌ 1 New Failure, 1 Pending, 2 Unrelated Failures

As of commit acf553f with merge base 2100527 (image):

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 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 15, 2025
@facebook-github-bot
Copy link
Contributor

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

Copy link

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 added a commit to Ninja91/executorch that referenced this pull request Aug 15, 2025
Summary:

- 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.

Note: Not verified with tosa reference model run.

Differential Revision: D80308822
@facebook-github-bot
Copy link
Contributor

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

Ninja91 added a commit to Ninja91/executorch that referenced this pull request Aug 15, 2025
Summary:
Pull Request resolved: pytorch#13448

- 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.

Note: Not verified with tosa reference model run.

Differential Revision: D80308822
@digantdesai digantdesai requested a review from per August 15, 2025 15:28
Ninja91 added a commit to Ninja91/executorch that referenced this pull request Aug 15, 2025
Summary:

- 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.

Note: Not verified with tosa reference model run.

Differential Revision: D80308822
@facebook-github-bot
Copy link
Contributor

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

Ninja91 added a commit to Ninja91/executorch that referenced this pull request Aug 15, 2025
Summary:
Pull Request resolved: pytorch#13448

- 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.

Note: Not verified with tosa reference model run.

Differential Revision: D80308822
@Ninja91 Ninja91 force-pushed the export-D80308822 branch 2 times, most recently from bad2472 to cd39120 Compare August 15, 2025 21:28
Ninja91 added a commit to Ninja91/executorch that referenced this pull request Aug 15, 2025
Summary:

- 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.

Note: Not verified with tosa reference model run.

Differential Revision: D80308822
@facebook-github-bot
Copy link
Contributor

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

@Ninja91
Copy link
Contributor Author

Ninja91 commented Aug 25, 2025

GH issue: #13635

Ninja91 added a commit to Ninja91/executorch that referenced this pull request Aug 25, 2025
Summary:
Pull Request resolved: pytorch#13448

- 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.

Reviewed By: digantdesai

Differential Revision: D80308822
Ninja91 added a commit to Ninja91/executorch that referenced this pull request Aug 25, 2025
Summary:

- 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.

Reviewed By: digantdesai

Differential Revision: D80308822
Ninja91 added a commit to Ninja91/executorch that referenced this pull request Aug 25, 2025
Summary:

- 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.

Reviewed By: digantdesai

Differential Revision: D80308822
@facebook-github-bot
Copy link
Contributor

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

1 similar comment
@facebook-github-bot
Copy link
Contributor

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

Ninja91 added a commit to Ninja91/executorch that referenced this pull request Aug 25, 2025
Summary:
Pull Request resolved: pytorch#13448

- 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.

Reviewed By: digantdesai

Differential Revision: D80308822
Ninja91 added a commit to Ninja91/executorch that referenced this pull request Aug 25, 2025
Summary:

- 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.

Reviewed By: digantdesai

Differential Revision: D80308822
Ninja91 added a commit to Ninja91/executorch that referenced this pull request Aug 25, 2025
Summary:

- 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.

Reviewed By: digantdesai

Differential Revision: D80308822
@facebook-github-bot
Copy link
Contributor

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

1 similar comment
@facebook-github-bot
Copy link
Contributor

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

Ninja91 added a commit to Ninja91/executorch that referenced this pull request Aug 25, 2025
Summary:
Pull Request resolved: pytorch#13448

- 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.

Reviewed By: digantdesai

Differential Revision: D80308822
…#13641)

Summary:

This diff implements a 16A8W (16-bit activations, 8-bit weights) quantization configuration utility for the ExecutorTorch ARM backend, following the feedback from D79746479.

## Key Changes

**1. New Quantization Configuration Function**
- Add `get_16a8w_quantization_config()` in `fbcode/executorch/backends/arm/quantizer/arm_quantizer.py`
- Provides 16-bit activations with HistogramObserver (better precision than 8A8W)
- Maintains 8-bit weights with MinMaxObserver/PerChannelMinMaxObserver (memory efficient)
- **Technically supported by TOSA through [EXT-INT16 extension/profile](https://www.mlplatform.org/tosa/tosa_spec.html#_conv2d)**

## Benefits
- **Better Precision**: 16-bit activations provide higher precision than 8-bit. Useful for carrying precision for recurring neural nets.
ghstack-source-id: 305459598
exported-using-ghexport

Reviewed By: 3l1

Differential Revision: D79763381
Summary:

- 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.

Reviewed By: digantdesai

Differential Revision: D80308822
@facebook-github-bot
Copy link
Contributor

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

@Ninja91
Copy link
Contributor Author

Ninja91 commented Aug 25, 2025

@pytorchmergebot

@Ninja91
Copy link
Contributor Author

Ninja91 commented Aug 25, 2025

@digantdesai @per strangely seeing https://github.com/pytorch/executorch/actions/runs/17219418500/job/48850992655 error which doesn't look related. Am i missing something here?

@Ninja91
Copy link
Contributor Author

Ninja91 commented Aug 25, 2025

Updated in a new PR: #13658. Closing this.

@Ninja91 Ninja91 closed this Aug 25, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. fb-exported partner: arm For backend delegation, kernels, demo, etc. from the 3rd-party partner, Arm
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants