from torch import randn
from torchmetrics.regression import R2Score
metric = R2Score()
metric.update(randn(10,), randn(10,))
fig_, ax_ = metric.plot()