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