How to multiply a vector by an array/matrix element-wise in numpy?

Question

I have a multidimensional array a whose shape is (32,3,5,5) and an array v with a shape of (32,). How could I multiply (i,3,5,5) with (i,) for each i using numpy other than a for-loop?

score 2 · Accepted Answer · edited Apr 29 '20 at 14:24

2

With a and v as the two arrays, few approaches could be suggested -

a*v[:,None,None,None]
a*v.reshape(-1, *[1]*3)
(a.T * v).T
np.einsum('i...,i->i...', a, v)

edited Apr 29 '20 at 14:24

smci

32,567
20
113
146

answered Jul 07 '17 at 18:01

Divakar

218,885
19
262
358

I tried the first approach you posted. It works! Thanks. – Xu Zhang Jul 07 '17 at 22:14

How to multiply a vector by an array/matrix element-wise in numpy?

1 Answers1