Skip to content

Conversation

@keyprocedure
Copy link
Contributor

@keyprocedure keyprocedure commented Mar 25, 2025

Summary

Fixes #9023
Prevents the partitioner from handling ops with mixed dtypes.

Test plan

Unable to directly test due to auto-casting of dtypes and existing dtype checks in verifier.py.

@pytorch-bot
Copy link

pytorch-bot bot commented Mar 25, 2025

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/9612

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

✅ No Failures

As of commit bc7dba2 with merge base 879b94f (image):
💚 Looks good so far! There are no failures yet. 💚

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

@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Mar 25, 2025
@keyprocedure keyprocedure changed the title Add mixed dtype check for XNNPACK partitioner (#9023) Add mixed dtype check for XNNPACK partitioner Mar 25, 2025
@keyprocedure
Copy link
Contributor Author

@pytorchbot label "topic: not user facing"

@mcr229
Copy link
Contributor

mcr229 commented Mar 26, 2025

do you mind running python -m unittest backends/xnnpack/test/ops/*.py

just to make sure there aren't any unintended consequences

@mcr229
Copy link
Contributor

mcr229 commented Mar 26, 2025

hmm seems like:
pull / unittest-buck / macos / macos-job (pull_request)

is passing which means all those tests are passing. Let's just try and fix the linter issues.

@mcr229
Copy link
Contributor

mcr229 commented Mar 27, 2025

will stamp and merge after all ci is run

@mcr229 mcr229 merged commit ec1cd04 into pytorch:main Mar 27, 2025
81 checks passed
@mcr229
Copy link
Contributor

mcr229 commented Mar 27, 2025

Thank you for your contribution!

kirklandsign pushed a commit that referenced this pull request Apr 11, 2025
### Summary
Fixes #9023
Prevents the partitioner from handling ops with mixed dtypes.

### Test plan
Unable to directly test due to auto-casting of dtypes and existing dtype
checks in verifier.py.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. topic: not user facing

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Partitioner to not delegate mixed dtype operators

3 participants