test: validate PT2 inference statistics in L0_torch_aoti

[mattwittwer]

What does the PR do?

Adds test_inference_statistics_sanity to qa/L0_torch_aoti/torch_aoti_infer_test.py. After running inferences, it fetches get_inference_statistics and asserts that the compute phase durations (compute_input + compute_infer + compute_output) never exceed the total successful request duration, for both per-request and per-batch statistics.

Checklist

• PR title reflects the change and is of format <commit_type>: <Title>
• Changes are described in the pull request.
• Related issues are referenced.
• Populated github labels field
• Added test plan and verified test passes.
• Verified that the PR passes existing CI.
• Verified copyright is correct on all changed files.
• Added succinct git squash message before merging ref.
• All template sections are filled out.
• Optional: Additional screenshots for behavior/output changes with before/after.

Commit Type:

Check the conventional commit type
box here and add the label to the github PR.

• build
• ci
• docs
• feat
• fix
• perf
• refactor
• revert
• style
• test

Related PRs:

Where should the reviewer start?

Test plan:

• CI Pipeline ID:

63772955

Caveats:

Background

Related Issues: (use one of the action keywords Closes / Fixes / Resolves / Relates to)

• closes GitHub issue: #xxx
  PT2/AOTI PyTorch backend reports incorrect compute_infer latency due to swapped ReportStatistics arguments #8874

[greptile-apps]

Greptile Summary

Adds an inference-statistics sanity test for the Torch AOTI model.

• Captures statistics before and after five requests so earlier suite traffic is excluded.
• Verifies that the summed input, inference, and output compute durations do not exceed successful-request duration for request and batch statistics.
• Adds targeted lint suppressions for imports following the common-path adjustment.

Confidence Score: 5/5

The PR appears safe to merge.

No blocking failure remains.

Important Files Changed

Filename	Overview
qa/L0_torch_aoti/torch_aoti_infer_test.py	The new baseline deltas and summed phase checks address both previously reported issues without leaving a concrete blocking failure.

_{Reviews (4): Last reviewed commit: "simplify comments" | Re-trigger Greptile}

[yinggeh]

LGTM. Although I think L0_response_statistics is a more suitable place.