Contents:
Examples
Utility functions for perturbation based predictions.
make_predictions(data, runner, batch_size)
make_predictions
Make predictions with the input data.
Process the data with the model runner and return the predictions.
data (np.ndarray) – An array of masked input data to be processed by the model.
runner (object) – An object that runs the model on the input data and returns predictions.
batch_size (int) – The number of masked inputs to process in each batch.
An array of predictions made by the model on the input data.
np.ndarray