I have a 1000 x 500 matrix. For each row, I want to calculate the dot product between the transpose of the row vector and the row vector itself (which should be a scalar), and sum over all rows. Right now I am using import numpy as np A = np.random.random((1000, 500)) res = sum(A.T @ A for i in range(A.shape[0])) Since this is the performance bottleneck of my algorithm, I wonder if there are faster ways to do this, peferably a Numpyic solution. Continue reading...
