Cortex_M backend: Add more model tests

[AdrianLundell]

Add model tests of currently not supported models

• yolo11
• wav2letter
• silero_vad

cc @digantdesai @freddan80 @per @zingo @oscarandersson8218 @mansnils @Sebastian-Larsson @robell @rascani

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/19720

• 📄 Preview Python docs built from this PR

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

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

• Run pull request jobs on OSDC runners in shadow mode

❌ 1 New Failure

As of commit e5e6b03 with merge base b4a9e72 ():

NEW FAILURE - The following job has failed:

• pull / android / run-emulator (gh)
  The process '/usr/bin/sh' failed with exit code 1

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

[Copilot]

Pull request overview

Adds new Cortex-M backend model tests intended to track lowering/acceleration progress for currently unsupported models.

Changes:

• Add YOLO11 dialect test using Ultralytics YOLO11n.
• Add Wav2Letter dialect test using the example torchaudio-based model.
• Add Silero VAD (16k) dialect test using the example model definition and fixed op-count expectations.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
backends/cortex_m/test/models/test_yolo11.py	Introduces a YOLO11n-based dialect test for progress tracking.
backends/cortex_m/test/models/test_wav2letter.py	Introduces a Wav2Letter dialect test using the examples model.
backends/cortex_m/test/models/test_silero_vad.py	Introduces a Silero VAD 16k dialect test with expected op-counts before/after transforms.

Comments suppressed due to low confidence (2)

backends/cortex_m/test/models/test_wav2letter.py:34

• The docstring says this model currently doesn’t lower due to missing conv1d support, but the test is not marked as expected-failure. As written, this will fail CI until conv1d is implemented. Consider using the parametrize(..., xfails={"wav2letter": "..."}, strict=False) mechanism (or an explicit pytest.mark.xfail) so the test can track progress without breaking the suite.

@parametrize("test_case", test_cases)
def test_dialect_wav2letter(test_case):
    """This model currently does largely not lower to accelerated kernels due to missing conv1d support, this test is to track development progress."""
    inputs = test_case.get_example_inputs()
    tester = CortexMTester(test_case.model, inputs)
    tester.test_dialect(
        ops_before_transforms,
        ops_after_transforms,
        qtol=10,
    )

backends/cortex_m/test/models/test_yolo11.py:45

• The docstring indicates YOLO11 currently does not lower in the Cortex-M backend, but the test isn’t marked xfail/skip. That will make the test suite fail until support lands. Use parametrize(..., xfails={"yolo11n": "..."}, strict=False) (or equivalent pytest.mark.xfail) so it can serve as a progress tracker without blocking CI.

@parametrize("test_case", test_cases)
def test_dialect_yolo11(test_case):
    """This model currently does not lower in the cortex-m backend, this test is to track development progress."""
    inputs = test_case.get_example_inputs()
    tester = CortexMTester(test_case.model, inputs)
    tester.test_dialect(
        ops_before_transforms,
        ops_after_transforms,
        qtol=10,
    )

---

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