Metal backend: fix linear_filter in test_modules

[manuelcandales]

No description provided.

[manuelcandales]

Stack from ghstack (oldest at bottom):

• Metal backend: preserve meta data in decompose_linear_pass #17237
• Metal backend: fail test_modules.py if torchao can't be imported #17236
• Metal backend: remove parakeet linear decomp #17116
• Metal backend: enable linear with bias #17115
• Metal backend: Add Metal int4 quantization support to Parakeet #17235
• -> Metal backend: fix linear_filter in test_modules #17253

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/17253

• 📄 Preview Python docs built from this PR

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

⏳ No Failures, 123 Pending

As of commit cf89a2b with merge base 20703d0 ():
💚 Looks good so far! There are no failures yet. 💚

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

[github-actions]

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.

[Copilot]

Pull request overview

This PR fixes a bug in the linear_filter function used for Metal backend int4 quantization. The old filter logic incorrectly validated that the weight dimension K must be divisible by the group_size (e.g., K % 32 == 0), but the Metal backend actually only requires K to be divisible by 8. This would have incorrectly prevented quantization of valid layers (e.g., layers with K=8 and group_size=32).

Changes:

• Fixed validation logic to check K % 8 instead of K % group_size
• Changed from silent filtering to explicit error handling with helpful error messages
• Added validation to reject group_size=0 early with a clear error message

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.