HoldoutIterator

class lightautoml.validation.base.HoldoutIterator(train, valid)[source]

Iterator for classic holdout - just predefined train and valid samples.

__init__(train, valid)[source]

Create iterator.

Parameters

get_validation_data()[source]

Just get validation sample.

apply_feature_pipeline(features_pipeline)[source]

Inplace apply features pipeline to iterator components.

Parameters: features_pipeline (FeaturesPipeline) – Features pipeline to apply.
Return type: HoldoutIterator
Returns: New iterator.

apply_selector(selector)[source]

Same as for basic class, but also apply to validation.

convert_to_holdout_iterator()[source]

Do nothing, just return itself.