Arm backend: Add fold_quantize option to tester

[tom-arm]

• Enables tests to be run with fold_quantize = False when required
• This is used when you do not want convert_pt2e to fold constants
• Temporary solution to enable testing of StaticCache in INT8

Change-Id: Ib25ea3949fc5f539c1a0a15565c3cbfe5099b9a1

cc @digantdesai @SS-JIA @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/18005

• 📄 Preview Python docs built from this PR

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

❌ 4 New Failures, 1 Pending

As of commit 949e14e with merge base 122fdef ():

NEW FAILURES - The following jobs have failed:

• pull / test-llama-runner-qnn-linux (fp32, qnn_8a8w, qnn) / linux-job (gh)
  RuntimeError: Command docker exec -t a1b32128cbde17aa68cfb9f5b6e1b85f87ecad7b24cd48a348a00efe6ad895d6 /exec failed with exit code 16
• pull / test-samsung-models-linux / linux-job (gh)
  RuntimeError: Command docker exec -t 587da83c93ca4e377c36de867ff17e2a437c10e64b89095374dcfec992c48613 /exec failed with exit code 1
• pull / test-samsung-quantmodels-linux / linux-job (gh)
  RuntimeError: Command docker exec -t 440b1c60d1378ff0eb075b71d1166f47e3733b99fc2dbce624054ce8deb3116e /exec failed with exit code 1
• trunk / test-qnn-model (fp32, ic4) / linux-job (gh)
  RuntimeError: Command docker exec -t bccecf375c7f4ee8299f5bc50228d9ea99f515869fb3a7ff111d97d6d1acd0bb /exec failed with exit code 1

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

[Copilot]

Pull request overview

This PR adds a fold_quantize option to the Arm backend test pipeline infrastructure, allowing tests to optionally disable constant folding during quantization (convert_pt2e). This is described as a temporary solution to enable testing of StaticCache in INT8 where you don't want convert_pt2e to fold constants.

Changes:

• Adds a fold_quantize parameter to ArmQuantize in quantize.py and threads it through to quantize_with_submodules
• Adds fold_quantize parameter to quantize_with_submodules in arm_quantizer.py and passes it to convert_pt2e
• Propagates the fold_quantize option through multiple pipeline classes in test_pipeline.py

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

File	Description
backends/arm/test/tester/quantize.py	Extends ArmQuantize with fold_quantize parameter, passes it to quantize_with_submodules
backends/arm/quantizer/arm_quantizer.py	Adds fold_quantize parameter to quantize_with_submodules, forwards to convert_pt2e
backends/arm/test/tester/test_pipeline.py	Threads fold_quantize through TosaPipelineINT, EthosUPipelineINTBase, EthosU55PipelineINT, EthosU85PipelineINT, QuantizationPipeline, and TosaPipelineMI pipeline classes

---

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

[zingo]

Maybe fix the copilot name change fold_quantze -> fold_quantize already now?

[tom-arm]

Maybe fix the copilot name change fold_quantze -> fold_quantize already now?

I have implemented CoPilot's suggestions, so the review is ready now

[tom-arm]

Failing jobs are unrelated