pytorch初学

• repeat

import torch
x = torch.tensor([1, 2, 3])
print(x)
print(x.repeat(4, 2))
print(x.repeat(4, 2, 1))

•  permute: 将对应维度的张量置换，有点numpy.transpose()的意思，具体可体会下面的例子

import torch

unpermuted=torch.tensor([[[1,2,3],[4,5,6]]])
print('unpermuted.size:',unpermuted.size())
print(unpermuted)

permuted=unpermuted.permute(2,0,1)
print('permuted.size:',permuted.size())
print(permuted)

•  view： 和numpy.reshape()类似

import torch

x1=torch.Tensor([[[1,2,3],[4,5,6]]])
x2=torch.Tensor([1,2,3,4,5,6])

print('x1.view:  ',x1.view(1,6))
print('x1.view:  ',x2.view(1,6))

x1=torch.Tensor([[[1,2,3],[4,5,6]]])
print('x1:\n',x1)
print('view:\n',x1.view(3,2))

print('permute:\n',x1.permute(0,2,1))

 

 

 若x1.vew(-1,2)表示reshape成一个2列的张量，多少行不确定。

• bmm：对存储在两个batch1和batch2内的矩阵进行批矩阵乘操作