import torch
from torchmetrics.image import TotalVariation
metric = TotalVariation()
values = [ ]
for _ in range(10):
    values.append(metric(torch.rand(5, 3, 28, 28)))
fig_, ax_ = metric.plot(values)