предотвратить разделение на ноль numpy
>>> a = np.array([-1, 0, 1, 2, 3], dtype=float)
>>> b = np.array([ 0, 0, 0, 2, 2], dtype=float)
# If you don't pass `out` the indices where (b == 0) will be uninitialized!
>>> c = np.divide(a, b, out=np.zeros(a.shape, dtype=float), where=b!=0)
>>> print(c)
[ 0. 0. 0. 1. 1.5]
Real Raccoon