import torch
from torchmetrics.audio import ScaleInvariantSignalDistortionRatio
target = torch.randn(5)
preds = torch.randn(5)
metric = ScaleInvariantSignalDistortionRatio()
metric.update(preds, target)
fig_, ax_ = metric.plot()