import torch
from torchmetrics import MetricCollection
from torchmetrics.classification import BinaryAccuracy, BinaryPrecision, BinaryRecall
metrics = MetricCollection([BinaryAccuracy(), BinaryPrecision(), BinaryRecall()])
values = []
for _ in range(10):
    values.append(metrics(torch.rand(10), torch.randint(2, (10,))))
fig_, ax_ = metrics.plot(values, together=True)