Skip to content

Avoid duplicating data in cross_val_score #7919

Closed
Closed
Avoid duplicating data in cross_val_score#7919
@ohadle

Description

@ohadle
ohadle
opened on Nov 21, 2016

I'm using a large dataset which takes up most of my memory. cross_val_score is basically unable to run on it (I get a MemoryError) since AFAICT data is duplicated for each fold (with list() here):

scikit-learn/sklearn/model_selection/_validation.py

Line 131 in 38f6a91

cv_iter = list(cv.split(X, y, groups))

Would it be possible to use views or otherwise avoid this duplication?

Metadata

Metadata

Assignees

No one assigned

    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