)
        >>> kendall = KendallRankCorrCoef(t_test=True, alternative='two-sided', num_outputs=2)
        >>> kendall(preds, target)
        (tensor([1., 1.]), tensor([nan, nan]))

    FNTg