I'm having trouble with powers with negative numbers: sometimes it gives me error, sometimes it does it without warnings or errors.
import numpy as np
FS = 1.28
X = (np.arange(20) / FS)[np.newaxis]
Y = np.real((X.T @ X) ** (-0.5))
for i in range (0, 20, 1):
for j in range (0, 20, 1):
print(Y[i][j])
This code doesn't give any problems.
This other one, instead, returns error:
W = np.real(((W @ W.T)**(-0.5)) @ W)
W is 2x128 matrix.
The error is:
W = np.real(((W @ W.T)**(-0.5)) @ W)
RuntimeWarning: invalid value encountered in power
in both cases I've made operations with a matrix and the same power, why in the second case it returns error ?