[ExecuTorch][WebGPU] cat op test suite (cases.py op-test framework)

[JulianCloudNTH]

Stack from ghstack (oldest at bottom):

• [ExecuTorch][WebGPU] aten.index.Tensor test suite (export + native golden) #20465
• [ExecuTorch][WebGPU] Add aten.index.Tensor (1D-self gather) #20464
• [ExecuTorch][WebGPU] Add clone op (aten.clone.default) #20463
• [ExecuTorch][WebGPU] Flatten landed-op test dirs to test/ops/test_<op>.py #20435
• -> [ExecuTorch][WebGPU] cat op test suite (cases.py op-test framework) #20399
• [ExecuTorch][WebGPU] Add cat op + ValueList graph support (aten.cat.default) #20398

Registers aten.cat.default in the cases.py op-test framework: a _cat_suite of 4 configs (3 inputs along dim 0, 2 inputs along dim 1, 3 inputs along dim 2, uneven split along dim 1) that generate_op_tests exports via VulkanPartitioner and compares to a torch golden on Dawn. Also adds test/ops/cat/test_cat.py (CatModule + CONFIGS + _op_delegated smoke test, with distinct per-input value ranges to catch cross-slab contamination) and the aten.cat.default partitioner-allowlist entry in tester.py.
@exported-using-ghexport

Differential Revision: D108793163

Differential Revision: D108793163

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/20399

• 📄 Preview Python docs built from this PR

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

⏳ No Failures, 77 Pending

As of commit 84a44d0 with merge base 200f64a ():
💚 Looks good so far! There are no failures yet. 💚

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

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[JulianCloudNTH]

@claude review

[claude]

Claude finished @JulianCloudNTH's task in 3m 25s —— View job

---

Review: WebGPU cat op-test suite

Read the full diff and compared it against the sibling op suites (permute, squeeze, mul) and the generator (generate_op_tests.py). This is a clean, well-scoped addition that follows the established op-test conventions closely. No blocking issues.

What's good

• CatModule + CONFIGS + smoke-test layout mirrors test_permute.py / test_squeeze.py exactly, and the cases.py registration matches the _squeeze_suite/_permute_suite shape (construct={"dim": dim}, golden_dtype="float32" with the copy-is-bit-identical rationale). Good consistency.
• This is the first variadic / multi-input op in the framework. CatModule.forward(self, *xs) + inputs=tuple(shapes) correctly exercises the generator's module(*inputs) / torch.export.export(module, inputs) path with multiple differently-shaped inputs — a genuinely new code path through the suite. The uneven config ((2,1,4)+(2,3,4)+(2,2,4) along dim 1) is a nice touch for catching per-input offset bugs.
• tester.py allowlist entry (aten.cat.default) is in the right place.

Minor notes (non-blocking)

1. The "distinct per-input value range" contamination trick only applies to the local smoke test, not the Dawn suite. _det_inputs (the base += 1000.0 ramps) is used by TestCat in test_cat.py, but the _cat_suite cases in cases.py pass bare shape tuples, so generate_op_tests._materialize produces them via torch.randn (all ~N(0,1), overlapping ranges). Cross-slab contamination is still detectable against the golden, but the explicit distinct-range property described in the PR body doesn't carry into the generated Dawn comparison. Worth a one-line clarification in the description, or consider an InputSpec(gen=...) ramp for one suite case if you want the strong signal on Dawn too.

2. _op_delegated's substring check is fragile (backends/webgpu/test/ops/cat/test_cat.py:60). "cat" in str(target) will also match scatter, concat, etc. Harmless here (a cat module has no such ops), and it's the same pattern as the siblings — but "cat" is a shorter, more collision-prone needle than "permute". If you ever generalize this helper, matching the full edge-op target (e.g. aten.cat) would be safer.

3. Helper duplication (_lower / _delegated / _op_delegated) is copy-pasted across every ops/*/test_*.py. This PR follows the existing convention, so it's fine — just flagging that this family of files is accumulating identical boilerplate that a shared _smoke_helpers.py could collapse later.

Note: I was unable to run pytest backends/webgpu/test/ops/cat/test_cat.py in this environment (command needed approval), so correctness here is by inspection only. The export/delegation logic is identical to the already-green sibling tests, so I'd expect it to pass.
· branch gh/JulianCloudNTH/48/head