Arm Backend: Generate input if not supplied in the non semihosting

[zingo]

This will generate fake input filled with 1 instead of no input if no input is supplied.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/5308

• 📄 Preview Python docs built from this PR

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

❌ 1 New Failure, 40 Pending

As of commit 2a6f91d with merge base fcbbef4 ():

NEW FAILURE - The following job has failed:

• trunk / test-models-macos (cmake, edsr, portable, macos-m1-stable, 90) / macos-job (gh)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 1

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

[zingo]

@pytorchbot label ciflow/trunk

[digantdesai]

sorry I don't see a fill, all I see is a check

[zingo]

In the normal case (non semihosting) the check always fails right now, as we don't have any input so the function exits too early by mistake. This check was added when we added the semihosting mode for the FVP to detect if the input file is wrong/missing. E.g. original this function created example input and filled it with 1 (that should be later in the function) but when we added semihosting we missed that we didn't run this code for non semihosting at all.

[digantdesai]

I think I get it. In Semihosting mode you don't want to fill 1's for some reason, right?

[facebook-github-bot]

@digantdesai has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[robell]

in semihosting there are files placed with smoketest input/output data so we can compare some more useful tensor input.
This assert was buggy for non-semihosting (run.sh flows) where it early exited the function so we wouldn't get 1's in the tensor, and so e.g. the delegated add would just report 0's instead of 2's.

[facebook-github-bot]

@digantdesai merged this pull request in 9c5994d.