ml_grid.model_classes.tabpfn_classifier_class

Defines the TabPFN Classifier model class.

Attributes

TABPFN_AVAILABLE

Classes

TabPFNClassifierClass

Initializes the TabPFNClassifierClass.

Module Contents

ml_grid.model_classes.tabpfn_classifier_class.TABPFN_AVAILABLE = True

class ml_grid.model_classes.tabpfn_classifier_class.TabPFNClassifierClass(parameter_space_size: str | None = None, model_version: str = 'v2.5_default', device: str = 'cpu', n_estimators: int = 4, subsample_samples: int | None = None, random_state: int = 42)

Bases: sklearn.base.BaseEstimator, sklearn.base.ClassifierMixin

Initializes the TabPFNClassifierClass.

Parameters:

• parameter_space_size (Optional[str]) – Size of the parameter space for optimization. Defaults to None.

• model_version (str) – The version of the TabPFN model to use.

• device (str) – The device to run the model on (‘cpu’ or ‘cuda’).

• n_estimators (int) – Number of ensemble members.

• subsample_samples (Optional[int]) – Subsample size for large datasets.

• random_state (int) – Random state for reproducibility.

Raises:

ImportError – If TabPFN is not installed.

model_version = 'v2.5_default'

device = 'cpu'

n_estimators = 4

subsample_samples = None

random_state = 42

parameter_space_size = None

algorithm_implementation

method_name: str = 'TabPFNClassifier'

parameter_vector_space: ml_grid.util.param_space.ParamSpace

parameter_space: Dict[str, Any]

fit(X: pandas.DataFrame, y: pandas.Series)

Fits the TabPFN model.

This method uses the hyperparameters set on the instance to create and fit the underlying TabPFNClassifier.

predict(X: pandas.DataFrame) → pandas.Series

Makes predictions using the fitted model.

predict_proba(X: pandas.DataFrame) → pandas.DataFrame

Returns probability estimates for predictions.