from torch import randn from torchmetrics.regression import TweedieDevianceScore metric = TweedieDevianceScore() metric.update(randn(10,), randn(10,)) fig_, ax_ = metric.plot()