本文共 4909 字，大约阅读时间需要 16 分钟。

LibTorch 张量操作与线性回归

在LibTorch中，张量操作与PyTorch相比，主要体现在基础操作的写法上。值得注意的首要事务是理解LibTorch与PyTorch之间的语法差异，尤其是变量命名方式的变化。

一 张量初始化

1. torch::zeros

#include 
   
    using namespace std;int main() {    torch::Tensor b = torch::zeros({2, 3});    cout << b << endl;    system("pause");    return 0;}
   

위키百科

2. torch::ones

#include 
   
    using namespace std;int main() {    torch::Tensor b = torch::ones({2, 3});    cout << b << endl;    system("pause");    return 0;}
   

위키百科

3. torch::eye

#include 
   
    using namespace std;int main() {    torch::Tensor b = torch::eye(3);    cout << b << endl;    system("pause");    return 0;}
   

위키百科

4. torch::full

#include 
   
    using namespace std;int main() {    torch::Tensor b = torch::full({2, 3}, 999);    cout << b << endl;    system("pause");    return 0;}
   

위키百科

5. torch::tensor

#include 
   
    using namespace std;int main() {    torch::Tensor  b = torch::tensor({        {1,2,3},        {4,5,6}    });    cout << b << endl;    system("pause");    return 0;}
   

위키百科

6. torch::rand

#include 
   
    using namespace std;int main() {    torch::Tensor b = torch::rand({2, 3});    cout << b << endl;    system("pause");    return 0;}
   

위키百科

7. torch::randn

#include 
   
    using namespace std;int main() {    torch::Tensor b = torch::randn({2, 3});    cout << b << endl;    system("pause");    return 0;}
   

위키百科

8. torch::randint

#include 
   
    using namespace std;int main() {    torch::Tensor b = torch::randint(1, 10, {2, 3});    cout << b << endl;    system("pause");    return 0;}
   

二 深浅拷贝

1. 浅拷贝

#include 
   
    using namespace std;int main() {    torch::Tensor b1 = torch::zeros({2, 3});    torch::Tensor b2 = b1;  // 浅拷贝    auto b3 = torch::Tensor(b1);}
   

2. 深拷贝

#include 
   
    using namespace std;int main() {    torch::Tensor b1 = torch::zeros({2, 3});    torch::Tensor b2 = b1.clone();    // 测试深拷贝后的独立性    b2[0][0] = 888;    cout << "b2: " << b2 << endl;    cout << "b1: " << b1 << endl;}
   

3. 张量大小一致的深拷贝

#include 
   
    using namespace std;int main() {    torch::Tensor b1 = torch::zeros({2, 3});    torch::Tensor b2 = torch::zeros_like(b1);    // 测试形状一致性    cout << "b2（ones-like）: " << b2 << endl;    b2 = torch::ones_like(b1);    cout << "b2（ones-like）：" << b2 << endl;    b2 = torch::rand_like(b1, torch::kFloat);    cout << "b2（random-like）：" << b2 << endl;}
   

三 C++ 常用数据类型转换为Tensor变量

1. 使用from_blob方法从数组转换Tensor

#include 
   
    using namespace std;int main() {    int a[] = {1, 2, 3, 4, 5, 6};    torch::Tensor b = torch::from_blob(a, {6}, torch::kInt);    cout << b << endl;}
   

2. 使用from_blob方法从浮点数数组转换Tensor

#include 
   
    using namespace std;int main() {    float data[] = {1, 2, 3, 4, 5, 6};    torch::Tensor f = torch::from_blob(data, {2, 3});}
   

3. 从std::vector

#include 
   
    using namespace std;int main() {    vector
    
      a = {1, 2, 3, 4, 5, 6};    // 方法一：直接使用 tensorSYSTEMrepr    torch::Tensor b = torch::tensor(a);    // 方法二：使用from_blob    torch::Tensor b = torch::from_blob(a.data(), {2}, torch::kFloat);}
    
   

四 张量维度变换

1. 拼接操作：torch::cat

#include 
   
    using namespace std;int main() {    auto a = torch::ones({2, 3});    auto a1 = torch::cat({a, a}, 0);    auto a2 = torch::cat({a, a}, 1);    cout << "a1: " << a1 << endl;    cout << "a2: " << a2 << endl;}
   

2. 堆叠操作：torch::stack

#include 
   
    using namespace std;int main() {    auto a = torch::ones({2, 3});    auto a1 = torch::stack({a, a}, 2);    cout << a1.shape() << endl;}
   

3. 切分操作：torch::chunk

#include 
   
    using namespace std;int main() {    auto a = torch::ones({2, 5});    auto chunks = torch::chunk(a, 2, 1);    for (int idx, t : chunks) {        cout << "idx: " << idx << " tensor: " << t << endl;    }}
   

4. unsqueeze操作（无维度展开）：torch::unsqueeze

#include 
   
    using namespace std;int main() {    auto a = torch::rand({2, 3});    // 查看原始形状    cout << a << endl;    // 增加一个维度（默认在第一个维度）    auto b = a.unsqueeze(1);    cout << b << endl;    // 测试形状变化    cout << b.shape() << endl;}
   

五 截取张量

1. 截取张量的特定部分

#include 
   
    using namespace std;int main() {    auto a = torch::rand({2, 3, 5, 4});    // 截取最后两个元素    auto b = a.slice(2, -2, -1);    cout << b << endl;}
   

六 张量的四则运算

1. 加法运算

#include 
   
    using namespace std;int main() {    auto a = torch::rand({2, 3});    auto b = torch::rand({3, 2});    auto c = a + b;    cout << c << endl;}
   

2. 乘法运算

#include 
   
    using namespace std;int main() {    auto a = torch::rand({2, 3});    auto b = torch::rand({3, 2});    auto c = torch::matmul(a, b);    cout << c << endl;}
   

3. 绝对值

#include 
   
    using namespace std;int main() {    auto a = torch::rand({2, 3});    auto b = torch::abs(a);    cout << b << endl;}
   

七 获取张量的大小

#include 
   
    using namespace std;int main() {    auto a = torch::rand({2, 3, 6, 6});    cout << a << endl;    cout << "Tensor size: " << a.sizes() << endl;}
   

八 其他操作（需cmd+T＞Code查看完整代码）

以上内容涵盖了LibTorch中常用操作的实现，对于线性回归部分，可参考标准教程或进一步优化模型结构。但请注意，在实际应用中，需根据需求选择适合的框架和模型结构。

转载地址：http://nwwfk.baihongyu.com/