Commit b8e000bd authored by Fabian Meyer's avatar Fabian Meyer
Browse files

Apply tests to new stepsize interface

parent caa0a41c

test/gdcpp.cpp

+9 −26
Original line number Diff line number Diff line
@@ -44,7 +44,7 @@ TEST_CASE("gradient_descent") 
        {

            GradientDescent<float,

                Paraboloid<float>,

                ConstantStepSize<float, Paraboloid<float>>,

                ConstantStepSize<float>,

                NoCallback<float>,

                ForwardDifferences<float, Paraboloid<float>>> optimizer;

            optimizer.setMaxIterations(100);
@@ -62,7 +62,7 @@ TEST_CASE("gradient_descent") 
        {

            GradientDescent<float,

                Paraboloid<float>,

                ConstantStepSize<float, Paraboloid<float>>,

                ConstantStepSize<float>,

                NoCallback<float>,

                BackwardDifferences<float, Paraboloid<float>>> optimizer;

            optimizer.setMaxIterations(100);
@@ -80,7 +80,7 @@ TEST_CASE("gradient_descent") 
        {

            GradientDescent<float,

                Paraboloid<float>,

                ConstantStepSize<float, Paraboloid<float>>,

                ConstantStepSize<float>,

                NoCallback<float>,

                CentralDifferences<float, Paraboloid<float>>> optimizer;

            optimizer.setMaxIterations(100);
@@ -98,7 +98,7 @@ TEST_CASE("gradient_descent") 
        {

            GradientDescent<float,

                Paraboloid<float>,

                ConstantStepSize<float, Paraboloid<float>>> optimizer;

                ConstantStepSize<float>> optimizer;

            optimizer.setMaxIterations(100);



            Vector xval(2);
@@ -110,13 +110,12 @@ TEST_CASE("gradient_descent") 
            REQUIRE_MATRIX_APPROX(xvalExp, result.xval, eps);

        }



        SECTION("limited change step")

        SECTION("Barzilai-Borwein step")

        {

            GradientDescent<float,

                Paraboloid<float>,

                LimitedChangeStep<float, Paraboloid<float>>> optimizer;

                BarzilaiBorweinStep<float>> optimizer;

            optimizer.setMaxIterations(100);

            // optimizer.setVerbose(true);



            Vector xval(2);

            xval << 2, 2;
@@ -127,27 +126,11 @@ TEST_CASE("gradient_descent") 
            REQUIRE_MATRIX_APPROX(xvalExp, result.xval, eps);

        }



        SECTION("barzilai borwein step")

        SECTION("Wolfe linesearch")

        {

            GradientDescent<float,

                Paraboloid<float>,

                BarzilaiBorweinStep<float, Paraboloid<float>>> optimizer;

            optimizer.setMaxIterations(100);



            Vector xval(2);

            xval << 2, 2;

            Vector xvalExp(2);

            xvalExp << 0, 0;



            auto result = optimizer.minimize(xval);

            REQUIRE_MATRIX_APPROX(xvalExp, result.xval, eps);

        }



        SECTION("Wolfe line search")

        {

            GradientDescent<float,

                Paraboloid<float>,

                WolfeLineSearch<float, Paraboloid<float>>> optimizer;

                WolfeBacktracking<float>> optimizer;

            optimizer.setMaxIterations(100);



            Vector xval(2);
@@ -163,7 +146,7 @@ TEST_CASE("gradient_descent") 
    SECTION("optimize Rosenbrock")

    {

        GradientDescent<float, Rosenbrock<float>,

            WolfeLineSearch<float, Rosenbrock<float>>> optimizer;

            WolfeBacktracking<float>> optimizer;

        optimizer.setMaxIterations(3000);

        optimizer.setMomentum(0.9);

        Vector xval(2);