How to calculate pairwise cosine distance of matrix using broadcasting in python

Question

I have a matrix like this where each row is a vector and need to calculate pairwise cosine distance between the vectors (in this example, as a 2x2 matrix) without using loops? I understand for cartesian distance one can pad up a dimension but don't know the best method here.

m = np.array([[1, 3, 5],[2,6,10]])

Answer 1

This type of question will get you berated with "Stack isn't for homework" but I guess you could just do this...

from numpy import linalg as LA
import numpy as np
m = np.array([[1, 3, 5],[2,6,10]])
cos = float(sum(m[0]*m[1]))/float(LA.norm(m[0])*LA.norm(m[1]))

of course if you wanted to avoid using linlag

float(sum(m[0]*m[1]))/(float(sum(m[0]*m[0]))**0.5*float(sum(m[1]*m[1]))**0.5)

See link -> https://en.wikipedia.org/wiki/Cosine_similarity