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