diff_values.py

import numpy as np


a = np.arange(0., 1.0, .1)
b = np.arange(0., 1.0, .1)

# modify 3 values to have differences
b[5:8] *= 1.1


indices_different = np.where(a != b)

# reason for indices_different[0] is because numpy is intrisically an n-dim
# linalg library so there is no such thing as 1-d arrays, and np.where returns
# a tuple of results, one entry per dimension in your input data. 
# Sucks a bit for 1-d arrays.
print("found {} different values between a and b. Indices: {}".format(len(indices_different[0]), indices_different[0]))
print("diff values in a {}".format(np.take(a, indices_different)))
print("diff values in b {}".format(np.take(b, indices_different)))