FIX skip array API tests when running with device="mps" without the PYTORCH_ENABLE_MPS_FALLBACK env var #27199
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
PCA uses parts of PyTorch that are not supported on the MPS device, this changes the estimator to raise an explicit exception with information on what to do.
|
I would rather not have pytorch specific things in the scikit-learn code base. I think the lower level error message returned by PyTorch is explicit enough, no? |
|
As discussed at the bi-weekly Array API meeting, it would be better to find a way to skip the Not 100% sure how to achieve that without adding more boilerplate though. EDIT: it should be easy to add a condition to the |
|
Good point. I've updated the code |
|
/cc @thomasjpfan for a quick review. |
|
Ooops I was wrong with the conflict resolution via the github UI. Let me push a fix quickly. |
|
It should be good now. |
thomasjpfan
approved these changes
Sep 7, 2023
REDVM
pushed a commit
to REDVM/scikit-learn
that referenced
this pull request
Nov 16, 2023
…YTORCH_ENABLE_MPS_FALLBACK env var (scikit-learn#27199) Co-authored-by: Olivier Grisel <olivier.grisel@ensta.org>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PCA uses parts of PyTorch that are not supported on the MPS device, this changes the estimator to raise an explicit exception with information on what to do.
I had to change the common Array API specific tests to handle the exception. There are more tests in the general common tests that would need adjusting. Not quite sure what to do.
sklearn/utils/tests/test_estimator_checks.py::test_check_estimator_clonessklearn/tests/test_common.py::test_estimators[PCA()-check_array_api_input(array_namespace=torch,dtype=float32,device=mps)]How is this handled for other cases where an estimator doesn't support a particular setup that is tested in the common tests?