permutation Importance warning scikit 0.25

[ThomasWolf0701]

When running permutation Importance I get the following warning:

C:\Users\Thomas Wolf\anaconda3\lib\site-packages\sklearn\utils\validation.py:70: FutureWarning: Pass classifier=False as keyword args. From version 0.25 passing these as positional arguments will result in an error
FutureWarning)
C:\Users\Thomas Wolf\anaconda3\lib\site-packages\sklearn\utils\validation.py:70: FutureWarning: Pass classifier=False as keyword args. From version 0.25 passing these as positional arguments will result in an error
FutureWarning)

[zeromh]

Same! Would love to see this fixed before it breaks.

[johncant]

Looks like this has been fixed in the latest master

[ThomasWolf0701]

Looks like this has been fixed in the latest master

Still looks the same as before to me. The pypi version is still the old one too.

[ThomasWolf0701]

cv = check_cv(self.cv, y, classifier=is_classifier(self.estimator)) seems to be the culprit in the permutation_importance.py function.

[ThomasWolf0701]

cv = check_cv(self.cv, y, is_classifier(self.estimator))
TypeError: check_cv() takes from 0 to 2 positional arguments but 3 were given

Is the error message with scikit-learn 1.0.2

[lopuhin]

This should be fixed in #48 , we're planning a PyPI release soon, closing this one as there are other issues open about sklearn compatibility, e.g. #42