-
-
Notifications
You must be signed in to change notification settings - Fork 25.8k
FIX Fix GridSearchCV regression in 1.5 with parameter grid with heterogeneous parameter values #29078
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
FIX Fix GridSearchCV regression in 1.5 with parameter grid with heterogeneous parameter values #29078
Conversation
I have added a test based on the original report. I also added a changelog, but there is definitely some room for improvement. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM!
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM. Thanks
Thanks! Ran into this today |
Arviz uses scipy.signal.gaussian, which was removed in 1.13. Most recent arviz uses scipy.signal.windows.gaussian scikit-learn 1.5.0 contained a regression (scikit-learn/scikit-learn#28352) that has been fixed in scikit-learn/scikit-learn#29078
* BLD: Fix broken versions Arviz uses scipy.signal.gaussian, which was removed in 1.13. Most recent arviz uses scipy.signal.windows.gaussian scikit-learn 1.5.0 contained a regression (scikit-learn/scikit-learn#28352) that has been fixed in scikit-learn/scikit-learn#29078 * BLD: Chasing errors, limit scipy/arviz in SBR To test the notebooks, need to install SBR extras, which included a version of arviz that isn't available on 3.9. Earlier version works, but restricts scipy version
…eous parameter values (scikit-learn#29078)
…eous parameter values (#29078)
* BLD: Fix broken versions Arviz uses scipy.signal.gaussian, which was removed in 1.13. Most recent arviz uses scipy.signal.windows.gaussian scikit-learn 1.5.0 contained a regression (scikit-learn/scikit-learn#28352) that has been fixed in scikit-learn/scikit-learn#29078 * BLD: Chasing errors, limit scipy/arviz in SBR To test the notebooks, need to install SBR extras, which included a version of arviz that isn't available on 3.9. Earlier version works, but restricts scipy version
Reference Issues/PRs
Fix #29074
What does this implement/fix? Explain your changes.
np.result_type
can raiseValueError
in some cases e.g.np.result_type(None, {'a': '1'})
. In that case we should use dtype object.I haven't added a test yet but will do soon.A test has been added.