Skip to content

[MRG] FIX row_norms return dtype same as input #12423

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

Closed

Conversation

jeremiedbb
Copy link
Member

Fixes #12422

Ensure that the norms returned by row_norms have the same dtype as their input, for dense and sparse input.

@jeremiedbb jeremiedbb closed this Oct 19, 2018
@jeremiedbb
Copy link
Member Author

Needs that k-means accept float32 to be fixed. It will be fixed in #11950

@rth
Copy link
Member

rth commented Oct 20, 2018

Needs that k-means accept float32 to be fixed.

But row_norms is used in other estimators, not just k-means, can't we fix it independently here?

@jeremiedbb
Copy link
Member Author

Well that's what I hoped but fixing this currently breaks k-means because it expects the dtype of the input norms to be float64. And I don't see the point to fix the k-means part here since it will involve parts of code that I removed in #11950. What's your opinion ?

@rth
Copy link
Member

rth commented Oct 22, 2018

It's up to you, we can also postpone it until your PR on K-Means is merged if you find it preferable..

@jeremiedbb
Copy link
Member Author

ok I'll check if it's an easy fix in kmeans :)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants