[WIP] Use Joblib backend hints rather than hardcoding #11345
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Uses the new prefer / require keywords from joblib/joblib#602. This allows users to control how jobs are parallelized in more situations. For example, training a RandomForest on a cluster of machines with the dask backend. Closes scikit-learn#8804
|
This pull request introduces 52 alerts and fixes 3 when merging 28260d9 into 93382cc - view on LGTM.com new alerts:
fixed alerts:
Comment posted by LGTM.com |
|
We just cherry-picked your commit in the |
|
Great, thanks. I'll close this PR then. If there are any comments on those changes I can make PRs against the #9486 branch. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Reference Issues/PRs
Closes #8804
See also #9486
See also joblib/joblib#602
This includes the changes from #602, since that's required for this. I wanted to get this up early, in the hopes it could be reviewed and included in the release (if possible). Currently, my changes are in 28260d9
What does this implement/fix? Explain your changes.
Uses joblib's new backend hints in every place where we currently hardcode the backend.
Currently (I think) only ForestClassifier / ForestRegressor rely on shared memory for correctness, and only in their
predictmethods. Everywhere else has been changed toprefer='threads'.Any other comments?
Open questions:
I'm not sure whether I should write similar tests for other backends.
non-default backend should indicate that in the
n_jobsparameterdescription. Would that be OK?
be using Dask's local cluster. If you all are OK with accepting Dask &
distributed as dependencies for the doc build then I'll get to work on that.
backend=, so that future PRs don't accidentally hardcode a backend?