Skip to content

[MRG] ENH: Add partial fit support to Pipeline#11321

Closed
stsievert wants to merge 4 commits intoscikit-learn:masterfrom
stsievert:pipeline-partial-fit
Closed

[MRG] ENH: Add partial fit support to Pipeline#11321
stsievert wants to merge 4 commits intoscikit-learn:masterfrom
stsievert:pipeline-partial-fit

Conversation

@stsievert
Copy link
Copy Markdown
Contributor

@stsievert stsievert commented Jun 19, 2018

Reference Issues/PRs

I'd like to see implemented for an adaptive parameter selection scheme that depends on partial_fit, dask/dask-searchcv#72.

This will also allow pipelines to work with dask_ml.Incremental, a class that wraps models that implement partial_fit.

This also resolves #3299.

What does this implement/fix? Explain your changes.

This implements Pipeline.partial_fit. It adds a new keyword argument partially_fit=[] to Pipeline.__init__. If changed from default it should take in a list of names to call partial_fit on.

Copy link
Copy Markdown
Member

@jnothman jnothman left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it will take time to reach consensus on this, so again, I don't think it fits into the current release.

def partial_fit(self, X, y, **fit_params):
"""Partially fit the model

Fit all the transforms one after the other and transform the
Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Are you sure we should be fitting all the transformers again on this sample??

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

Labels

API Needs Decision Requires decision

Projects

None yet

Development

Successfully merging this pull request may close these issues.

partial_fit support for Pipeline

3 participants