torch.reshape(input, shape)函数使用举例

参考链接: torch.reshape(input, shape)

说明:

返回一个张量,内容和原来的张量相同,但是具有不同形状.
并且尽可能返回视图,否则才会返回拷贝,
因此,需要注意内存共享问题.
传入的参数可以有一个-1,
表示其具体值由其他维度信息和元素总个数推断出来.

代码实验:

(base) PS C:\Users\chenxuqi> python
Python 3.7.4 (default, Aug  9 2019, 18:34:13) [MSC v.1915 64 bit (AMD64)] :: Anaconda, Inc. on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import torch
>>> a = torch.arange(4.)
>>> a
tensor([0., 1., 2., 3.])
>>> torch.reshape(a, (2, 2))
tensor([[0., 1.],
        [2., 3.]])
>>> b = torch.tensor([[0, 1], [2, 3]])
>>> b
tensor([[0, 1],
        [2, 3]])
>>> torch.reshape(b, (-1,))
tensor([0, 1, 2, 3])
>>>
>>>
>>>
>>> a = torch.arange(12)
>>> a
tensor([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11])
>>> a = torch.reshape(a,(3,4))
>>> a
tensor([[ 0,  1,  2,  3],
        [ 4,  5,  6,  7],
        [ 8,  9, 10, 11]])
>>>
>>>
>>>
>>>
>>>
>>>
>>> a = torch.arange(12)
>>> b = a.reshape(3,4)
>>> a
tensor([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11])
>>> b
tensor([[ 0,  1,  2,  3],
        [ 4,  5,  6,  7],
        [ 8,  9, 10, 11]])
>>> a[0] = 2020
>>> b[0][1] = 20200910
>>> a
tensor([    2020, 20200910,        2,        3,        4,        5,        6,
               7,        8,        9,       10,       11])
>>> b
tensor([[    2020, 20200910,        2,        3],
        [       4,        5,        6,        7],
        [       8,        9,       10,       11]])
>>> # 由此可见,两者共享内存
...
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> a = torch.arange(12)
>>> b = a.reshape(3,4)
>>> c = b.permute(1,0).reshape(2,6)  # permute(1,0)返回的张量不连续
>>> a
tensor([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11])
>>> b
tensor([[ 0,  1,  2,  3],
        [ 4,  5,  6,  7],
        [ 8,  9, 10, 11]])
>>> c
tensor([[ 0,  4,  8,  1,  5,  9],
        [ 2,  6, 10,  3,  7, 11]])
>>>
>>>
>>> a[0] = 9999
>>> b[0][1] = 8888
>>> c[1][5] = 7777
>>> a
tensor([9999, 8888,    2,    3,    4,    5,    6,    7,    8,    9,   10,   11])
>>> b
tensor([[9999, 8888,    2,    3],
        [   4,    5,    6,    7],
        [   8,    9,   10,   11]])
>>> c
tensor([[   0,    4,    8,    1,    5,    9],
        [   2,    6,   10,    3,    7, 7777]])
>>>
>>> # 由此可见a和b共享内存,但是a和b不与c共享内存
...
>>>
>>>
>>>