tabensemb.model.PytorchTabular._bayes_eval

method

PytorchTabular._bayes_eval(model, X_train, y_train, X_val, y_val)

Evaluate the model for Bayesian optimization iterations. The larger one of the training loss and the validation loss is returned by default.

Parameters:

model

The model returned by _new_model().

X_train

The training data from _train_data_preprocess().

y_train

The target of the training data from _train_data_preprocess().

X_val

The validation data from _train_data_preprocess().

y_val

The target of the validation data from _train_data_preprocess().

Returns:

float

The metric of the Bayesian hyperparameter optimization iteration.