import torch from torchmetrics.classification import MultilabelAUROC metric = MultilabelAUROC(num_labels=3) metric.update(torch.rand(20,3), torch.randint(2, (20,3))) fig_, ax_ = metric.plot()