python 矩阵切片

假设n是一个numpy或者torch.tensor张量,那么

n[a:b,c:d]则代表从a到b行(不含b行),c到d列(不含d列)的切片

当然,ab、cd中的数是可以省略掉,只要abcd中有一个数即可

import torch
m = [[-1.5805,  1.6342,  0.2004,  0.1923,  0.2503, -0.0467, -1.9203], \
[-1.6402,  0.5139,  1.8963,  0.5725,  0.4209,  0.6523, -1.6669], \
[-1.8716,  0.0378,  0.5280,  1.3742,  0.5176,  0.3444, -1.4203], \
[-1.7557,  0.1472,  0.0563,  0.5518,  1.3115,  0.0915, -1.7161], \
[-1.9358,  0.1463, -0.2262,  0.3781,  0.2074,  1.5654, -1.4256]]
n = torch.tensor(a)
print(n)
print(n[1:4,0:2])
print(n[:4,0:2])
print(n[1:4,:2])
print(n[1:,])