Support multiple axes in `Tensor` reduce methods

[Haydn Vestal]

i.e. min, max, sum, product

Example:

# ...

@tc.post
def average(tensor, *axes):
    assert hasattr(axes, "__len__")

    if axes:
        return tensor.sum(axes) / functools.reduce(operator.mul, tensor.shape[x] for x in axes)
    else:
        return tensor.sum() / tensor.size