pytorch中有索引方法吗



我最近在学习pytorch。但是这个问题太奇怪了。。

x=np.arrage(24)
ft=torch.FloatTensor(x)
print(floatT.view([@1])[@2])

答案=张量([[13.,16.],[19.,22.]](

是否存在满足答案的索引方法@1和@2?

如果您首先获取您关心的值,然后使用view将其解释为矩阵,则会更容易思考:

# setting up
>>> import torch
>>> import numpy as np
>>> x=np.arange(24) + 3 # just to visualize the difference between indices and values
>>> x
array([ 3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19,
20, 21, 22, 23, 24, 25, 26])
# taking the values you want and viewing as matrix
>>> ft = torch.FloatTensor(x)
>>> ft[[13, 16, 19, 22]]
tensor([16., 19., 22., 25.])
>>> ft[[13, 16, 19, 22]].view(2,2)
tensor([[16., 19.],
[22., 25.]])

通过viewft作为具有6列的张量:

ft.view(-1, 6)
Out[]:
tensor([[ 0.,  1.,  2.,  3.,  4.,  5.],
[ 6.,  7.,  8.,  9., 10., 11.],
[12., 13., 14., 15., 16., 17.],
[18., 19., 20., 21., 22., 23.]])

将元素(1319(和(1622(放置在彼此之上。现在您只需要从右侧的行/列中提取它们:

.view(-1, 6)[2:, (1, 4)]
Out[]:
tensor([[13., 16.],
[19., 22.]])

最新更新