Print matrix diagonals method

I need to print (or store) all diagonals of a matrix. This includes the main diagonal ( { (i,i) | i = 1, 2, ... min(m,n) } ). Then all super and sub diagonals.

For example:

1  2  3  4
 5  6  7  8
 9 10 11 12
13 14 15 16

The diagonals should be (order doesn't matter):

[
    [1, 6, 11, 16],
    [2, 7, 12],
    [5, 10, 15],
    [3, 8],
    [9, 14],
    [4],
    [13]
]

I believe I have a decent solution, but if there's a simpler / more efficient way to do this let me know. Input is assumed to be a list of lists as a NumPy array already has some methods to do this.

def print_diags(Matrix):
    m, n = len(Matrix), len(Matrix[0])
    for offset in range(min(m,n)):
        diag_upper = [row[i + offset]
                      for i, row in enumerate(Matrix)
                      if 0 <= i + offset < n]
        if offset != 0:
            diag_lower = [row[i - offset]
                      for i, row in enumerate(matrix)
                      if 0 <= i - offset < m]
            print(diag_upper, diag_lower)
        else:
            print(diag_upper)

Here's a generator in a generator:

m = [[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12], [13, 14, 15, 16]]

def diags(mat):
    width, height = len(mat[0]), len(mat)
    def diag(sx, sy):
        for x, y in zip(range(sx, height), range(sy, width)):
            yield mat[x][y]
    for sx in range(height):
        yield list(diag(sx, 0))
    for sy in range(1, width):
        yield list(diag(0, sy))

Usage:

>>> list(diags(m))
[[1, 6, 11, 16], [5, 10, 15], [9, 14], [13], [2, 7, 12], [3, 8], [4]]

diag starts from a coordinate and walks down the diagonal, yielding elements from it. It doesn't raise an IndexError, because zip exits after the shortest iterable is exhausted.

The last four lines are a bit inelegant, but I don't know a better way.

StackExchange Code Review Q#141714, answer score: 2