torch.flatten()

最新推荐文章于 2024-03-02 18:39:23 发布

Pr Young

最新推荐文章于 2024-03-02 18:39:23 发布

阅读量244

点赞数

分类专栏：深度学习文章标签：深度学习

本文链接：https://blog.csdn.net/weixin_47414034/article/details/127227425

版权

深度学习专栏收录该内容

35 篇文章 4 订阅

订阅专栏

将张量拉成一维的向量

x=torch.randn(2,3,2)
x2=torch.flatten(x,0)
x3=torch.flatten(x,1)
x4=torch.flatten(x,2)

import torch
x=torch.randn(2,3,2)
print(x)

#生成：
[    [   [-0.5829,  0.8214],
         [ 0.6218,  0.3298],
         [ 0.0222, -0.8473]   ],

     [   [ 0.1044, -1.8784],
         [ 1.2323,  2.6551],
         [ 0.0382,  0.6649]   ]    ]


x2=torch.flatten(x,0)#等价于x2=torch.flatten(x)
print(x2)
[-0.5829,  0.8214,  0.6218,  0.3298,  0.0222, -0.8473,  0.1044, -1.8784, 1.2323,  2.6551,0.0382,  0.6649]

import torch
x=torch.randn(2,3,2)
print(x)

#生成：
[    [   [-0.5829,  0.8214],
         [ 0.6218,  0.3298],
         [ 0.0222, -0.8473]   ],

     [   [ 0.1044, -1.8784],
         [ 1.2323,  2.6551],
         [ 0.0382,  0.6649]   ]    ]



x3=torch.flatten(x,1)
print(x3)


[   [-0.5829,  0.8214,  0.6218,  0.3298,  0.0222, -0.8473],
    [ 0.1044, -1.8784,  1.2323,  2.6551,  0.0382,  0.6649]    ]

import torch
x=torch.randn(2,3,2)
print(x)

#生成：
[    [   [-0.5829,  0.8214],
         [ 0.6218,  0.3298],
         [ 0.0222, -0.8473]   ],

     [   [ 0.1044, -1.8784],
         [ 1.2323,  2.6551],
         [ 0.0382,  0.6649]   ]    ]


x4=torch.flatten(x,2)
print(x4)



[    [   [-0.5829,  0.8214],
         [ 0.6218,  0.3298],
         [ 0.0222, -0.8473]     ],

     [   [ 0.1044, -1.8784],
         [ 1.2323,  2.6551],
         [ 0.0382,  0.6649]     ]    ]