Arm backend: Rework reporting of qspecs and qconfigs

[martinlsm]

This PR contains four commits:

1. Move quantizer_reporter out from cortex_m.quantizer
   When importing backends.cortex_m.quantizer.quantizer_reporter from
   backends.arm.quantizer.arm_quantizer_utils, a cyclic dependency chain is
   is formed. The problem is that quantizer_reporter triggers
   backends/cortex_m/quantizer/__init__.py when imported, which in turn has
   imports leading back to the Arm backend. To fix this problem, move
   quantizer_reporter to backends/cortex_m so it can be imported without
   forming any cycle.

2. Arm backend: Remove _QuantizerReporterUserMixin
   _QuantizerReporterUserMixin was a duplicated class with the name
   QuantizerReporterUser. Remove the former to instead use the latter.

3. Arm backend: Add label attribute to QuantizationConfig
   The quantizer reporter logs the quantization config in a human-readable
   format. Prior to this patch, this was done with the help of a dict
   called SUPPORTED_QCONFIGS, which was defined in quantizer_reporter.py
   and populated by the user. This patch reworks this concept by instead
   adding a label attribute to QuantizationConfig that the reporter can
   use to print the config in a human-readable format.

4. Arm backend: Rework reporting of qspecs
   The quantization reporter prints quantization specs in human-readable
   format. Prior to this patch, this was implemented such that
   quantizer_reporter.py defined a dict SUPPORTED_QSPECS which was
   populated by the user. This dict would map qspec objects to string
   representations. This patch removes this dict and instead modifies
   the helper function _qspec_repr to return a compact string
   representation based on the attributes of the qspec.

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

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/19016

• 📄 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, 6 Unrelated Failures

As of commit 3476cd8 with merge base e281726 ():

NEW FAILURE - The following job has failed:

• pull / unittest / macos / macos-job (gh)
  ninja: build stopped: subcommand failed

FLAKY - The following job failed but was likely due to flakiness present on trunk:

• pull / test-qnn-testsuite-linux / test-backend-linux (qnn, models) / linux-job (gh) (matched linux rule in flaky-rules.json)
  The runner has received a shutdown signal. This can happen when the runner service is stopped, or a manually started runner is canceled.

BROKEN TRUNK - The following jobs failed but was present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• pull / unittest / windows / windows-job (gh) (trunk failure)
  ##[error]The operation was canceled.
• pull / unittest-editable / macos / macos-job (gh) (trunk failure)
  export/tests/test_target_recipes.py::TestTargetRecipes::test_mv2_model
• pull / unittest-editable / windows / windows-job (gh) (trunk failure)
  ##[error]The operation was canceled.
• trunk / unittest-release / macos / macos-job (gh) (trunk failure)
  export/tests/test_target_recipes.py::TestTargetRecipes::test_mv2_model
• trunk / unittest-release / windows / windows-job (gh) (trunk failure)
  ##[error]The operation was canceled.

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

[martinlsm]

@pytorchbot label ciflow/trunk

[martinlsm]

@pytorchbot label "partner: arm"

[martinlsm]

@pytorchbot label "release notes: none"

[Copilot]

Pull request overview

This PR refactors quantization reporting for Arm/Cortex-M backends to avoid import cycles and to make qconfig/qspec reporting self-describing (via QuantizationConfig.label and attribute-based qspec formatting) rather than relying on external registration dicts.

Changes:

• Move quantizer_reporter to backends/cortex_m/ and update imports/build targets to eliminate a Cortex-M ↔ Arm cyclic dependency.
• Replace qconfig/qspec “registry dict” reporting with QuantizationConfig.label and a new qspec_repr() helper for compact, attribute-based qspec formatting.
• Update Arm/Cortex-M quantizers and tests to use the new reporter API (QuantizerInfo.qconfig_label) and new string representations.

Reviewed changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
backends/cortex_m/test/misc/test_quantizer_reporter.py	Updates tests to use the relocated reporter module and asserts new qspec string formatting via qspec_repr().
backends/cortex_m/quantizer_reporter.py	Relocates and refactors reporter: removes qspec/qconfig registries, adds qspec_repr(), and changes QuantizerInfo to use qconfig_label.
backends/cortex_m/quantizer/quantizer.py	Updates reporter import path to the new module location.
backends/cortex_m/quantizer/quantization_configs.py	Stops registering qspec/qconfig dicts; instead sets human-readable labels directly on config instances.
backends/cortex_m/quantizer/TARGETS	Removes per-subdir reporter target; depends on the new top-level //executorch/backends/cortex_m:quantizer_reporter.
backends/cortex_m/TARGETS	Adds a new Buck target exporting quantizer_reporter.py from the Cortex-M root.
backends/arm/quantizer/quantization_config.py	Adds optional label field to QuantizationConfig for debug/reporting.
backends/arm/quantizer/arm_quantizer_utils.py	Removes duplicated reporter mixin and switches to QuantizerReporterUser; updates qconfig labeling logic.
backends/arm/quantizer/arm_quantizer.py	Constructs and propagates meaningful label strings when creating quantization configs (instead of external registration).

---

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