import torch
from torchmetrics.nominal import TschuprowsT
metric = TschuprowsT(num_classes=5)
values = [ ]
for _ in range(10):
    values.append(metric(torch.randint(0, 4, (100,)), torch.randint(0, 4, (100,))))
fig_, ax_ = metric.plot(values)