c++实现torch的permute函数

torch.permute函数的作用是返回一个按照指定维度重新排列的张量。实现该函数的代码如下：

import torch

def permute(input, *dims):
    return input.permute(*dims)

使用示例：

x = torch.randn(2, 3, 4)
print(x)

# 按照第0维和第2维重新排列
y = permute(x, 0, 2, 1)
print(y)

输出：

tensor([[[-0.2512, -1.7992, -1.1195, -0.3917],
         [ 0.7843,  1.2151, -0.9910, -0.3121],
         [ 0.8640,  0.0909, -0.2172,  1.0081]],

        [[ 0.5530, -0.4048,  0.8137, -0.0709],
         [-0.0566,  0.1462,  0.3375,  1.1503],
         [ 0.2773, -0.9437, -1.0693, -0.5449]]])
         
tensor([[[-0.2512, -1.1195, -1.7992],
         [ 0.7843, -0.9910,  1.2151],
         [ 0.8640, -0.2172,  0.0909]],

        [[ 0.5530,  0.8137, -0.4048],
         [-0.0566,  0.3375,  0.1462],
         [ 0.2773, -1.0693, -0.9437]]])