Arm backend: Update _adjust_weight_qspec_for_conv_transpose

[mansnils]

• FusedMovingAvgObsFakeQuantize only supports channel axis 0, fall back to non-fused (MovingAveragePerChannelMinMaxObserver + FakeQuantize) when applicable.
• Always check/update ch_axis for FakeQuantize/observer constructor regardless if ch axis is correct in Qspec.
• Adds unit tests.

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

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/18094

• 📄 Preview Python docs built from this PR

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

❌ 1 New Failure, 1 Cancelled Job

As of commit b2e6040 with merge base 569e7a9 ():

NEW FAILURE - The following job has failed:

• Lint / link-check / lint-urls (gh)
  Process completed with exit code 1.

CANCELLED JOB - The following job was cancelled. Please retry:

• trunk / test-mcu-cortex-m-backend / linux-job (gh)
  ##[error]The operation was canceled.

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

[Copilot]

Pull request overview

This PR updates the Arm backend’s conv_transpose2d weight-qspec adjustment to correctly handle per-channel axis selection (including TorchAO ctor rebuilding) and adds unit tests validating the behavior for both PTQ and QAT flows.

Changes:

• Update _adjust_weight_qspec_for_conv_transpose to always fix ch_axis (including rebuilding QAT/PTQ observer/fake-quant ctors) and to fall back from fused fake-quant when axis != 0.
• Add targeted unit tests for axis correction and fused→non-fused fallback behavior in conv_transpose2d quantization preparation.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
backends/arm/quantizer/quantization_annotator.py	Adjusts conv_transpose2d weight qspec axis handling and rebuilds TorchAO constructors (with fused fallback behavior).
backends/arm/test/ops/test_transpose_conv2d.py	Adds unit tests covering QAT/PTQ axis correction and fused/non-fused fake-quant selection.

---

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