Implementation:Scikit learn Scikit learn EnableHalvingSearchCV

Knowledge Sources	Scikit_learn Scikit-learn Docs
Domains	Machine Learning, Hyperparameter Tuning
Last Updated	2026-02-08 15:00 GMT

Overview

Concrete tool for enabling the experimental Successive Halving search-estimators in scikit-learn's model_selection module.

Description

Importing this module dynamically sets HalvingRandomSearchCV and HalvingGridSearchCV as attributes of the sklearn.model_selection module. These are experimental estimators that implement the Successive Halving strategy for hyperparameter search, which progressively eliminates poor candidates using increasing amounts of resources.

Usage

Import this module before importing HalvingRandomSearchCV or HalvingGridSearchCV from sklearn.model_selection. This is required because these estimators are experimental and their API may change without deprecation cycles.

Code Reference

Source Location

• Repository: scikit-learn
• File: sklearn/experimental/enable_halving_search_cv.py

Signature

# Module-level side effects:
from sklearn.model_selection._search_successive_halving import (
    HalvingGridSearchCV,
    HalvingRandomSearchCV,
)
setattr(model_selection, "HalvingRandomSearchCV", HalvingRandomSearchCV)
setattr(model_selection, "HalvingGridSearchCV", HalvingGridSearchCV)

Import

from sklearn.experimental import enable_halving_search_cv  # noqa
from sklearn.model_selection import HalvingRandomSearchCV, HalvingGridSearchCV

I/O Contract

Inputs

Name	Type	Required	Description
(none)	N/A	N/A	This is a side-effect module; importing it enables the experimental estimators

Outputs

Name	Type	Description
HalvingRandomSearchCV	class	Experimental successive halving random search estimator added to model_selection
HalvingGridSearchCV	class	Experimental successive halving grid search estimator added to model_selection

Usage Examples

Basic Usage

from sklearn.experimental import enable_halving_search_cv  # noqa
from sklearn.model_selection import HalvingGridSearchCV
from sklearn.ensemble import RandomForestClassifier

clf = RandomForestClassifier(random_state=0)
param_grid = {"max_depth": [3, 5, 10], "min_samples_split": [2, 5, 10]}
search = HalvingGridSearchCV(clf, param_grid, random_state=0)

Related Pages

• Principle:Scikit_learn_Scikit_learn_Grid_Search