Write golden input artifacts in the tensor's own byte order

[rascani]

Summary

_dump_golden_artifacts wrote inputs through contiguous(), which means contiguous_format and re-lays a channels_last tensor back to NCHW. The .bin then disagrees with the dim_order the accompanying .pte declares, and the runtime ingests it as a raw memcpy, so anything replaying a golden input against the program feeds it transposed data.

Permuting by dim_order() first writes the bytes the runtime consumes. It is a no-op for a contiguous tensor, which is what every flow in the suite feeds today: all of its inputs are torch.randn and friends, and the permutes and transposes are inside the models rather than in the inputs.

Outputs deliberately keep contiguous(). reference_output is the eager result, so a model ending in a permute hands back a view whose dim_order describes the source rather than the result, while the program materializes that output contiguously. Permuting there would write the pre-permute values and corrupt the goldens for the permute and transpose operator tests, which is what the new test for that case pins down.

Test plan

Nothing consumes the channels_last path yet. The Cortex-M flow is the first to hand the harness such inputs, and CI uploads these artifacts, so this is a prerequisite for it.

Authored with Claude Code.

[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.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21851

• 📄 Preview Python docs built from this PR

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

⏳ No Failures, 6 Pending

As of commit 9597db3 with merge base 5c3cb65 ():
💚 Looks good so far! There are no failures yet. 💚

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