Skip to content

update scikit-learn dependency? #749

Closed

Description

@amueller

The current tests fail with later scikit-learn releases in several places, because the code relies on classes and attributes that have been removed.
I would suggest using 0.20 from Sept 2018 as minimum version.
That introduced the current ColumnTransformer and SimpleImputer and OneHotEncoder and so you'd be able to write current sklearn.

Activity

  1. mfeurer commented on Jul 24, 2019

    @mfeurer
    Collaborator

    I would suggest using 0.20 from Sept 2018 as minimum version.

    We had a longer discussion at the Paris workshop on this topic and came to the conclusion that we try to allow multiple versions where possible. We should discuss how to proceed at the next workshop.

    Anyway, I totally agree that we also need to test 0.21.X.

  2. amueller commented on Jul 24, 2019

    @amueller
    ContributorAuthor

    Ok I should really work on that CZI grant but happy to look into this soon (cause I'm too lazy to have an extra environment to run openml tests ;)

  3. mfeurer commented on Jul 26, 2019

    @mfeurer
    Collaborator

    Closed via #752.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions