127
    >>> a = np.array([1, 2, 3], dtype=np.float32)
    >>> ma.minimum_fill_value(a)
    inf

    rŒ