torch.cat() stack()函数使用说明，含实例及运行结果

先放张图，如果能看明白就不需要看下面实例解释

torch.cat() 函数

torch.cat() 函数定义

torch.cat(tensors, dim=0, *, out=None) → Tensor

参数及功能

** tensors** (sequence of Tensors) – 任何相同类型张量的 Python序列
dim (int, optional) – 拼接维度，默认为0

二维数据实例解释

参数dim=0

import torch

T1 = torch.tensor([[1, 2, 3],
                   [4, 5, 6],
                   [7, 8, 9]])
T2 = torch.tensor([[10, 20, 30],
                   [40, 50, 60],
                   [70, 80, 90]])

print(T1.shape)
print("=============================================")
print(torch.cat((T1, T2), dim=0).shape)
print("=============================================")
print(torch.cat((T1, T2), dim=0))

结果为

torch.Size([3, 3])
=============================================
torch.Size([6, 3])
=============================================
tensor([[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9],
        [10, 20, 30],
        [40, 50, 60],
        [70, 80, 90]])

可以看到是沿负y轴方向进行拼接

参数dim=1

import torch

T1 = torch.tensor([[1, 2, 3],
                   [4, 5, 6],
                   [7, 8, 9]])
T2 = torch.tensor([[10, 20, 30],
                   [40, 50, 60],
                   [70, 80, 90]])

print(T1.shape)
print("=============================================")
print(torch.cat((T1, T2), dim=1).shape)
print("=============================================")
print(torch.cat((T1, T2), dim=1))

结果为

torch.Size([3, 3])
=============================================
torch.Size([3, 6])
=============================================
tensor([[ 1,  2,  3, 10, 20, 30],
        [ 4,  5,  6, 40, 50, 60],
        [ 7,  8,  9, 70, 80, 90]])

可以看到是沿X轴正方向进行拼接

参数dim=-1

与dim=1相同

torch.stack() 函数

torch.stack() 函数定义

torch.stack(tensors, dim=0, *, out=None) → Tensor

参数及功能

tensors (sequence of Tensors) – 要连接的张量序列
dim (int) – 拼接维度，默认为0

二维数据实例解释

参数dim=0

import torch

T1 = torch.tensor([[1, 2, 3],
                   [4, 5, 6],
                   [7, 8, 9]])

T2 = torch.tensor([[10, 20, 30],
                   [40, 50, 60],
                   [70, 80, 90]])
print(torch.stack((T1, T2), dim=0).shape)
print("=============================================")
print(torch.stack((T1, T2), dim=0))

结果为

torch.Size([2, 3, 3])
=============================================
tensor([[[ 1,  2,  3],
         [ 4,  5,  6],
         [ 7,  8,  9]],

        [[10, 20, 30],
         [40, 50, 60],
         [70, 80, 90]]])

这个是按3维中的第0维拼接，可参考最上面图

参数dim=1

import torch

T1 = torch.tensor([[1, 2, 3],
                   [4, 5, 6],
                   [7, 8, 9]])

T2 = torch.tensor([[10, 20, 30],
                   [40, 50, 60],
                   [70, 80, 90]])
print(torch.stack((T1, T2), dim=1).shape)
print("=============================================")
print(torch.stack((T1, T2), dim=1))

结果为

torch.Size([3, 2, 3])
=============================================
tensor([[[ 1,  2,  3],
         [10, 20, 30]],

        [[ 4,  5,  6],
         [40, 50, 60]],

        [[ 7,  8,  9],
         [70, 80, 90]]])

参数dim=2

import torch

T1 = torch.tensor([[1, 2, 3],
                   [4, 5, 6],
                   [7, 8, 9]])

T2 = torch.tensor([[10, 20, 30],
                   [40, 50, 60],
                   [70, 80, 90]])
print(torch.stack((T1, T2), dim=2).shape)
print("=============================================")
print(torch.stack((T1, T2), dim=2))

结果为

torch.Size([3, 3, 2])
=============================================
tensor([[[ 1, 10],
         [ 2, 20],
         [ 3, 30]],

        [[ 4, 40],
         [ 5, 50],
         [ 6, 60]],

        [[ 7, 70],
         [ 8, 80],
         [ 9, 90]]])

参数dim=-1

结果与dim=2相同

参考博文及感谢

部分内容参考以下链接，这里表示感谢 Thanks♪(･ω･)ﾉ
参考博文1 torch.cat() 官方文档
https://pytorch.org/docs/1.13/generated/torch.cat.html
参考博文2 torch.stack()官方文档
https://pytorch.org/docs/1.13/generated/torch.stack.html
参考博文3 torch.cat()函数的官方解释，详解以及例子
https://blog.csdn.net/xinjieyuan/article/details/105208352
参考博文4 torch.stack()的官方解释，详解以及例子
https://blog.csdn.net/xinjieyuan/article/details/105205326