Keep pruning cross-validation folds stable across epochs - #318
Merged
Merged
Conversation
|
This pull request has not seen any recent activity. |
not522
added
bug
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
Pruning keeps a fitted estimator for each CV fold across epochs, but recreates the folds on every epoch. A shuffled splitter with a mutable
RandomStatethen puts previously trained samples into the validation fold, contaminating its score.Description of the changes
Materialize the splits once per trial and reuse them for its epochs. This preserves per-trial splitter behavior while keeping each estimator's held-out samples fixed.
The regression trains actual
SGDClassifierinstances and asserts that validation samples never appear in their accumulated training samples, across three epochs and two trials. It fails on the unchanged base. Local sklearn integration suite: 30 passed. Changed-file Black, isort, flake8 and mypy checks passed. Other integration suites were not run.AI assistance was used for implementation and local verification.
Fixes #317.