tabensemb.model.WideDeep._pred_single_model#

method

WideDeep._pred_single_model(model, X_test, verbose, **kwargs)[source]#

Predict using the model trained in _train_single_model().

Parameters:
model:

The model returned by _new_model() and trained in _train_single_model().

X_test:

The data from _data_preprocess() or _train_data_preprocess().

verbose:

Verbosity.

**kwargs:

Parameters to train the model returned by _get_params(). It contains all arguments in _initial_values().

Returns:
np.ndarray

Prediction of the target.

Notes

For deep learning models with mini-batch training (dataloaders), if an AbstractWrapper will be used to extract hidden representations, the batch_size when inferring should be the length of the dataset. See tabensemb.model.PytorchTabular._pred_single_model() and tabensemb.model.WideDeep._pred_single_model().