LBFGS++

https://github.com/yixuan/lbfgspp
Admin
A header-only C++ library that implements the Limited-memory BFGS algorithm (L-BFGS) for...
Tokens:2,452
Snippets:21
Trust Score:9.7
License:MIT
Update:5 months ago
Show doc for...
Tokens:
Raw
### Compile and Run LBFGS++ Example

Source: https://github.com/yixuan/lbfgspp/blob/master/README.md

Compile the C++ example using g++ with specified include paths for Eigen and LBFGS++. Execute the compiled binary to see the optimization results.

```bash
$ g++ -I/path/to/eigen -I/path/to/lbfgspp/include -O2 example.cpp
$ ./a.out
23 iterations
x =
1 1 1 1 1 1 1 1 1 1
f(x) = 1.87948e-19

```

--------------------------------

### Build C++ Examples with CMake

Source: https://github.com/yixuan/lbfgspp/blob/master/examples/CMakeLists.txt

This snippet iterates through a list of C++ source files, creates an executable for each, sets the C++ standard to 17, and links the necessary libraries.

```cmake
foreach (source example-quadratic.cpp example-rosenbrock-box.cpp example-rosenbrock-bracketing.cpp example-rosenbrock-comparison.cpp example-rosenbrock.cpp)
    get_filename_component(example ${source} NAME_WLE)
    add_executable(${example} ${source})
    set_property(TARGET ${example} PROPERTY CXX_STANDARD 17)
    target_link_libraries(${example} PRIVATE lbfgspp Eigen3::Eigen)
endforeach ()
```

--------------------------------

### Box-Constrained Rosenbrock Example

Source: https://github.com/yixuan/lbfgspp/blob/master/README.md

Demonstrates using LBFGSBSolver for optimization with simple bounds on parameters. Requires the LBFGSB.h header.

```cpp
#include <Eigen/Core>
#include <iostream>
#include <LBFGSB.h>  // Note the different header file

using Eigen::VectorXd;
using namespace LBFGSpp;

class Rosenbrock
{
private:
    int n;
public:
    Rosenbrock(int n_) : n(n_) {}
    double operator()(const VectorXd& x, VectorXd& grad)
    {
        double fx = 0.0;
        for(int i = 0; i < n; i += 2)
        {
            double t1 = 1.0 - x[i];
            double t2 = 10 * (x[i + 1] - x[i] * x[i]);
            grad[i + 1] = 20 * t2;
            grad[i]     = -2.0 * (x[i] * grad[i + 1] + t1);
            fx += t1 * t1 + t2 * t2;
        }
        return fx;
    }
};

int main()
{
    const int n = 10;
    // Set up parameters
    LBFGSBParam<double> param;  // New parameter class
    param.epsilon = 1e-6;
    param.max_iterations = 100;

    // Create solver and function object
    LBFGSBSolver<double> solver(param);  // New solver class
    Rosenbrock fun(n);

    // Bounds
    VectorXd lb = VectorXd::Constant(n, 2.0);
    VectorXd ub = VectorXd::Constant(n, 4.0);

    // Initial guess
    VectorXd x = VectorXd::Constant(n, 3.0);

    // x will be overwritten to be the best point found
    double fx;
    int niter = solver.minimize(fun, x, fx, lb, ub);

    std::cout << niter << " iterations" << std::endl;
    std::cout << "x = \n" << x.transpose() << std::endl;
    std::cout << "f(x) = " << fx << std::endl;

    return 0;
}

```

--------------------------------

### Run L-BFGS Minimization with Default Line Search

Source: https://github.com/yixuan/lbfgspp/blob/master/README.md

Set up L-BFGS parameters, create a solver instance, provide an initial guess, and run the minimization. The initial guess vector `x` will be overwritten with the optimal solution.

```cpp
int main()
{
    const int n = 10;
    // Set up parameters
    LBFGSParam<double> param;
    param.epsilon = 1e-6;
    param.max_iterations = 100;

    // Create solver and function object
    LBFGSSolver<double> solver(param);
    Rosenbrock fun(n);

    // Initial guess
    VectorXd x = VectorXd::Zero(n);
    // x will be overwritten to be the best point found
    double fx;
    int niter = solver.minimize(fun, x, fx);

    std::cout << niter << " iterations" << std::endl;
    std::cout << "x = \n" << x.transpose() << std::endl;
    std::cout << "f(x) = " << fx << std::endl;

    return 0;
}

```

--------------------------------

### Run L-BFGS Minimization with Bracketing Line Search

Source: https://github.com/yixuan/lbfgspp/blob/master/README.md

Configure and run the LBFGS++ solver using a specific line search algorithm, 'LineSearchBracketing', by providing it as a template parameter to LBFGSSolver.

```cpp
int main()
{
    const int n = 10;
    // Set up parameters
    LBFGSParam<double> param;
    param.epsilon = 1e-6;
    param.max_iterations = 100;

    // Create solver and function object
    LBFGSSolver<double, LineSearchBracketing> solver(param);
    Rosenbrock fun(n);

    // Initial guess
    VectorXd x = VectorXd::Zero(n);
    // x will be overwritten to be the best point found
    double fx;
    int niter = solver.minimize(fun, x, fx);

    std::cout << niter << " iterations" << std::endl;
    std::cout << "x = \n" << x.transpose() << std::endl;
    std::cout << "f(x) = " << fx << std::endl;

    return 0;
}

```

--------------------------------

### Link LBFGS++ in BUILD.bazel

Source: https://github.com/yixuan/lbfgspp/blob/master/README.md

Specifies the LBFGS++ library as a dependency for a C++ library target within a Bazel BUILD.bazel file.

```bzl
# BUILD.bazel

cc_library(
    name = "my_lib",
    srcs = ["my_lib.cc"],
    hdrs = ["my_lib.h"],
    deps = ["@lbfgspp"],
)

```

--------------------------------

### Add LBFGS++ Bazel Dependency

Source: https://github.com/yixuan/lbfgspp/blob/master/README.md

Integrates LBFGS++ into a Bazel project by adding the library as a dependency in the MODULE.bazel file. Automatically resolves transitive dependencies on Eigen.

```bzl
# MODULE.bazel

#...
bazel_dep(name = "lbfgspp", version = "4.0.0")
git_override(
    module_name = "lbfgspp",
    commit = "c524a407fb85b74807f53de5a3ca2ddbcc164e54",
    remote = "https://github.com/yixuan/LBFGSpp.git",
)
# ...

```

--------------------------------

### Define Rosenbrock Function for Minimization

Source: https://github.com/yixuan/lbfgspp/blob/master/README.md

Implement the Rosenbrock function and its gradient for use with the LBFGS++ solver. This functor must return the objective function value and overwrite a gradient vector.

```cpp
#include <Eigen/Core>
#include <iostream>
#include <LBFGS.h>

using Eigen::VectorXd;
using namespace LBFGSpp;

class Rosenbrock
{
private:
    int n;
public:
    Rosenbrock(int n_) : n(n_) {}
    double operator()(const VectorXd& x, VectorXd& grad)
    {
        double fx = 0.0;
        for(int i = 0; i < n; i += 2)
        {
            double t1 = 1.0 - x[i];
            double t2 = 10 * (x[i + 1] - x[i] * x[i]);
            grad[i + 1] = 20 * t2;
            grad[i]     = -2.0 * (x[i] * grad[i + 1] + t1);
            fx += t1 * t1 + t2 * t2;
        }
        return fx;
    }
};

```
Tokens:
Raw