probly.evaluation.ood.out_of_distribution_detection_aupr¶

probly.evaluation.ood.out_of_distribution_detection_aupr(in_distribution: ndarray, out_distribution: ndarray) → float[source]¶

Perform out-of-distribution detection using AUPR.

This metric evaluates how well the model distinguishes between in- and out-of-distribution samples, focusing more on positive class (OOD) precision and recall.

Parameters:

in_distribution – In-distribution prediction functionals.
out_distribution – Out-of-distribution prediction functionals.

Returns:

Area under the precision-recall curve.