Arm backend: Test prequantized NFRU weights

[Michiel-Olieslagers]

Load the published NFRU int8 checkpoint and restore its legacy PT2E parameters and buffers. Convert it without calibration or requantization, then exercise the resulting graph through the TOSA and VGF test pipelines.

Authored with assistance from Codex.

Change-Id: I8c0b1f0998d578cc32e92d50c8d5beeeb33e411c

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/23237

• 📄 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:

• Long queue for g5 runnners

❌ 1 Cancelled Job, 3 Unrelated Failures

As of commit 3ff3408 with merge base 2f78245 ():

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

• trunk / test-samsung-models-linux / linux-job (gh)
  ##[error]received SIGINT

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

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

• pull / test-arm-backend-no-driver (test_pytest_models_tosa) / linux-job (gh) (trunk failure)
  RuntimeError: Command bash /__w/_temp/exec_script failed with exit code 1
• trunk / test-arm-backend-vkml (test_pytest_models_vkml) / linux-job (gh) (trunk failure)
  RuntimeError: Command bash /__w/_temp/exec_script failed with exit code 1
• Windows MSVC Build / build-windows-arm64 (gh) (trunk failure)

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

[zingo]

Will this fix/adress some of the GitHub CI problem we got in #23182 ?
See the tests in the PR or https://hud.pytorch.org/hud/pytorch/executorch/main/1?per_page=50&name_filter=test-arm for more info.

[zingo]

Update:

@psiddh tried to fix the problem during the night and manage to fix 2/3 problems
#23265
#23268
Left after this are merged seem to only be a few numerical/high "mean abs error" problems.