Fix check_decision_proba_consistency random failure #19225
Merged
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
Apparently it was necessary for the Gradient Boosting classifiers on 32 bit linux... So apparently the goal is to create deterministic ties rather than close but non-deterministic non-ties at the rounding error level. I will revert this part of the change. |
thomasjpfan
approved these changes
Jan 21, 2021
There was a problem hiding this comment.
LGTM
Side note: Rounding to support different platforms may become a pattern: #19221
ogrisel
commented
Jan 21, 2021
adrinjalali
approved these changes
Jan 23, 2021
glemaitre
pushed a commit
to glemaitre/scikit-learn
that referenced
this pull request
Apr 22, 2021
* FIX more deterministic check_decision_proba_consistency * Trigger [cd build] * Re-add rounding * Trigger [cd build] * Avoid redundant phrasing in comment [ci skip]
glemaitre
pushed a commit
that referenced
this pull request
Apr 28, 2021
* FIX more deterministic check_decision_proba_consistency * Trigger [cd build] * Re-add rounding * Trigger [cd build] * Avoid redundant phrasing in comment [ci skip]
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Outdated suggestions cannot be applied.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.
Tentative fix for #19224.
X_testwas generated from the RNG singleton ofnp.randomand therefore the test was not deterministic.X_testfrom the same bloby distribution as the training set which is less likely to result in samples close to the decision boundary although I am not sure this has any actual impact. But it seems more natural to do so.I removed the rounding thingy that I did not understand. How rounding could possibly reduce the likelihood of ties?Anyway lets see how the CI likes this including the
[cd build].