reject.utils

Module for utils.

Module Contents

Functions

compute_correct(→ numpy.typing.NDArray)

Compute correct predictions.

aggregate_preds(→ tuple[numpy.typing.NDArray, ...)

Aggregate predictions to get stack, mean, and label.

generate_synthetic_output(...)

Generate synthetic NN output for showcasing functions.
reject.utils.compute_correct(y_true: numpy.typing.NDArray, y_pred: numpy.typing.NDArray) → numpy.typing.NDArray[source]

Compute correct predictions.

Parameters:

  • y_true (NDArray) – Array of true labels. Shape (n_observations,).

  • y_pred (NDArray) – Array of predictions. Shape (n_observations, n_classes) or (n_observations, n_samples, n_classes).

Returns:

Array of correct predictions. Shape (n_observations,).

Return type:

NDArray

Raises:

ValueError – If shape of y_pred or y_true is invalid.

reject.utils.aggregate_preds(y_pred: numpy.typing.NDArray) → tuple[numpy.typing.NDArray, numpy.typing.NDArray, numpy.typing.NDArray][source]

Aggregate predictions to get stack, mean, and label.

Parameters:

y_pred (NDArray) – Array of predictions. Shape (n_observations, n_classes) or (n_observations, n_samples, n_classes).

Returns:

Stack (rank 2 or 3), mean (rank 2), and label (rank 1) of predictions.

Return type:

tuple[NDArray, NDArray, NDArray]

reject.utils.generate_synthetic_output(num_samples: int, num_observations: int, concat: bool = True) → tuple[numpy.typing.NDArray, numpy.typing.NDArray] | tuple[tuple[numpy.typing.NDArray, numpy.typing.NDArray], tuple[numpy.typing.NDArray, numpy.typing.NDArray]][source]

Generate synthetic NN output for showcasing functions.

Parameters:

  • num_samples (int) – Number of samples to draw per observation.

  • num_observations (int) – Number of observations.

  • concat (bool, optional) – Whether to concatenate ID and OOD samples, by default True.

Returns:

  • Union[tuple[NDArray, NDArray], tuple[tuple[NDArray, NDArray],

  • tuple[NDArray, NDArray]]] – Tuple of synthetic predictions and true labels.