histogram_oods#

scio.eval.histogram_oods(conf_ind, conf_oods, *, oods_title=None, score_and_layers=None, **hist_kw)[source]#

For a given score, plot histograms over all OoD sets.

Parameters:

conf_ind (NDArray) – Confidence scores of In-Distribution samples. Shape (n_ind_samples,).
conf_oods (tuple[NDArray, ...]) – Same but for iterable of Out-of-Distribution samples. Shapes (n_oodi_samples,).
oods_title (tuple[str, ...], optional) – See compute_confidence(). Used only for legend purposes.
score_and_layers (ScoreClassifAndLayers | str, optional) –
The score (and layers) used to compute the confidence scores. Example for ScoreClassifAndLayers:
```
score_and_layers = KNN(k=6), [(1, 1)]
```
Used only for legend purposes.
**hist_kw – Passed to sns.histplot().

Returns:

ax (plt.Axes) – The matplotlib axes containing the plot.