Sampling matrix

[Weixin-An]

How is the sampling matrix obtained? which algorithm is used? Thanks a lot!

[usanzhu]

hi，have you sovled this problem？I’m confused with it, too. If solved, could you please help me with it, thanks!!

[Guaishou74851]

The following code provides a method of generating an M-by-N random Gaussian matrix:

import numpy as np
from scipy.linalg import orth

M = 512
N = 1024
Phi = orth(np.random.randn(N, N))[:M]

print(Phi @ Phi.T)

[usanzhu]

The following code provides a method of generating an M-by-N random Gaussian matrix:

import numpy as np
from scipy.linalg import orth

M = 512
N = 1024
Phi = orth(np.random.randn(N, N))[:M]

print(Phi @ Phi.T)

thanks a lot!!

[huyanjia]

以下代码提供了一种生成 M×N 随机高斯矩阵的方法：

你好，采用此代码随机生成的高斯矩阵需要正交化或者其他变化吗，这样训练之后重建的效果好吗

[huyanjia]

The following code provides a method of generating an M-by-N random Gaussian matrix:

import numpy as np
from scipy.linalg import orth

M = 512
N = 1024
Phi = orth(np.random.randn(N, N))[:M]

print(Phi @ Phi.T)

Hello! Do the randomly generated Gaussian matrices from this code need to be orthogonalized or otherwise modified? Will the reconstruction quality be good after training?

[huyanjia]

The following code provides a method of generating an M-by-N random Gaussian matrix:

import numpy as np
from scipy.linalg import orth

M = 512
N = 1024
Phi = orth(np.random.randn(N, N))[:M]

print(Phi @ Phi.T)

Hello! Do the randomly generated Gaussian matrices from this code need to be orthogonalized or otherwise modified? Will the reconstruction quality be good after training?