设计师值得拥有的设计导航_个人网站制作的选题意义_长沙百度开户_什么是网站优化

1. 举例：

在深度学习过程中，我们经常会用到掩码矩阵，比如我们有一个矩阵A表示如下,希望得到矩阵B，在矩阵A的非零位置表示为1，零位置表示为0，
$$\begin{array}{c}A=\left[\begin{array}{ccccc}2& 3& 0& 0& 0\\ & & & & \\ 1& 3& 4& 2& 0\end{array}\right]\to B=\left[\begin{array}{ccccc}1& 1& 0& 0& 0\\ & & & & \\ 1& 1& 1& 1& 0\end{array}\right]\end{array}$$

2. python 代码

#!/usr/bin/env python
# -*- coding:utf-8 -*-
# @FileName  :MatrixNoneZero2OnesLike.py
# @Time      :2024/12/22 16:20
# @Author    :Jason Zhang
import torch
import torch.nn as nnclass MatrixNoneZero2OnesLike(object):def __init__(self, in_matrix):self.in_matrix = in_matrixself._result = torch.zeros_like(self.in_matrix)@propertydef result(self):my_result = torch.zeros_like(self.in_matrix)my_result_bool = self.in_matrix.to(torch.bool)self._result = my_result.masked_fill(my_result_bool, 1)return self._resultif __name__ == "__main__":run_code = 0matrix_mya = torch.tensor([[2, 3, 0, 0, 0], [1, 3, 4, 2, 0]])my_matrix_test = MatrixNoneZero2OnesLike(matrix_mya)my_matrix_test_result = my_matrix_test.resultprint(f"matrix_mya=\n{matrix_mya}")print(f"my_matrix_test_result=\n{my_matrix_test_result}")

• 结果：

matrix_mya=
tensor([[2, 3, 0, 0, 0],[1, 3, 4, 2, 0]])
my_matrix_test_result=
tensor([[1, 1, 0, 0, 0],[1, 1, 1, 1, 0]])

3. 邻接矩阵

$$\begin{array}{c}B=\left[\begin{array}{ccccc}1& 1& 0& 0& 0\\ & & & & \\ 1& 1& 1& 1& 0\end{array}\right]\to C=\left[\begin{array}{c}\left[\begin{array}{ccccc}1& 1& 0& 0& 0\\ & & & & \\ 1& 1& 0& 0& 0\\ & & & & \\ 0& 0& 0& 0& 0\\ & & & & \\ 0& 0& 0& 0& 0\\ & & & & \\ 0& 0& 0& 0& 0\end{array}\right]\\ \\ \left[\begin{array}{ccccc}1& 1& 1& 1& 0\\ & & & & \\ 1& 1& 1& 1& 0\\ & & & & \\ 1& 1& 1& 1& 0\\ & & & & \\ 0& 0& 0& 0& 0\\ & & & & \\ 0& 0& 0& 0& 0\end{array}\right]\end{array}\right]\end{array}$$

• python code

#!/usr/bin/env python
# -*- coding:utf-8 -*-
# @FileName  :MatrixNoneZero2OnesLike.py
# @Time      :2024/12/22 16:20
# @Author    :Jason Zhang
import torch
import torch.nn as nnclass MatrixNoneZero2OnesLike(object):def __init__(self, in_matrix):self.in_matrix = in_matrixself._result = torch.zeros_like(self.in_matrix)@propertydef result(self):my_result = torch.zeros_like(self.in_matrix)my_result_bool = self.in_matrix.to(torch.bool)self._result = my_result.masked_fill(my_result_bool, 1)return self._resultclass AdjacencyMatrix(object):def __init__(self, in_matrix):self.in_matrix = in_matrixself.row, self.column = self.in_matrix.shapeself._result = torch.zeros((self.row, self.column, self.column))@propertydef result(self):self._result = torch.zeros((self.row, self.column, self.column))matrix_dim1 = torch.unsqueeze(self.in_matrix, dim=1)matrix_bmm = torch.bmm(matrix_dim1.transpose(-1, -2), matrix_dim1)self._result = matrix_bmmreturn self._resultif __name__ == "__main__":run_code = 0matrix_mya = torch.tensor([[2, 3, 0, 0, 0], [1, 3, 4, 2, 0]])my_matrix_test = MatrixNoneZero2OnesLike(matrix_mya)my_matrix_test_result = my_matrix_test.resultmy_adjacency = AdjacencyMatrix(my_matrix_test_result)my_adjacency_result = my_adjacency.resultprint(f"matrix_mya=\n{matrix_mya}")print(f"my_matrix_test_result=\n{my_matrix_test_result}")print(f"my_adjacency_result=\n{my_adjacency_result}")

• 结果：

matrix_mya=
tensor([[2, 3, 0, 0, 0],[1, 3, 4, 2, 0]])
my_matrix_test_result=
tensor([[1, 1, 0, 0, 0],[1, 1, 1, 1, 0]])
my_adjacency_result=
tensor([[[1, 1, 0, 0, 0],[1, 1, 0, 0, 0],[0, 0, 0, 0, 0],[0, 0, 0, 0, 0],[0, 0, 0, 0, 0]],[[1, 1, 1, 1, 0],[1, 1, 1, 1, 0],[1, 1, 1, 1, 0],[1, 1, 1, 1, 0],[0, 0, 0, 0, 0]]])