Update MemoryPlanning Verifier to only assume model has user input if it has at least one tensor input

[jhelsby]

Fixes #10522

Summary

MemoryPlanning verifier currently blows up if all the user inputs are prims. #10522 suggested:

We need to improve the logic of the _do_user_inputs_exist to probably just return false if all the inputs are prims.

This PR implements this suggestion, with accompanying unit tests. _do_user_inputs_exist now returns True if it has at least one tensor input, and False otherwise.

Test plan

Added unit tests to test_memory_planning.py and ran them with

pytest exir/tests/test_memory_planning.py::test_memory_planning.py

Please note you must comment out the line 69 from /pytest.ini for this to work:

--ignore=exir/tests/test_memory_planning.py

On my machine, I also had to comment out line 60 from test_memory_planning.py for the test to run without errors.

torch.ops.load_library("//executorch/kernels/portable:custom_ops_generated_lib")

[The tests I wrote aren't dependent on this library.]

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/10617

• 📄 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:

• VolumeLimitExceeded Issue for linux.2xlarge and linux.4xlarge

❌ 1 New Failure

As of commit 8a87967 with merge base 9d899c9 ():

NEW FAILURE - The following job has failed:

• pull / android / run-emulator (gh)
  The process '/usr/bin/sh' failed with exit code 255

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

[pytorch-bot]

Didn't find following labels among repository labels: Release notes: exir

[jhelsby]

@pytorchbot label "release notes: exir"

[digantdesai]

Started CI (Need to do for new, external contributions). Rebase please.

Also marked it as Ready to signal reviewers that it's, well, ready for a review. Thanks for the contribution. Welcome to Executorch 😊

[jhelsby]

Thank you, I just rebased. I also fixed the linting issues which the CI tests identified.

[Sorry, I messed this up, hence the subsequent force-push.]

[digantdesai]

cc @JacobSzwejbka can you take a look? 🙏

[larryliu0820]

@JacobSzwejbka gentle ping

[JacobSzwejbka]

I dont see any obvious issue with the code, the all red ci though made me think it was still under iteration.

[jhelsby]

Many thanks for looking at this @JacobSzwejbka. Is there anything I can do now to help get this PR closed?

[larryliu0820]

Kicked off CI jobs.

[larryliu0820]

I don't think there's any other comment so approving

[jhelsby]

I'm so sorry @larryliu0820, I accidentally merged the branch which seems to have stopped the running workflows. Could you please kick off the CI jobs again?

[jhelsby]

Please correct me if I'm wrong, but I believe the failing check (pull / android / run-emulator (pull_request)) is unrelated to my change. It seems to be failing in other PRs too. For example:

#12010

has the same failed check here, with the same failure output:

FAILURES!!!
Tests run: 17,  Failures: 16

What is the best way to proceed to get this merged?

@larryliu0820

[larryliu0820]

Yeah unrelated. Merging. Thanks for the PR!

[jhelsby]

Fantastic, thank you so much for your help and patience throughout! This was my first code contribution to a major open-source project. 😄