main.cpp: function fitting

main.cpp

@@ -1,9 +1,54 @@
 #include "robbie.hpp"
 #include <fmt/format.h>
+#include <fmt/ostream.h>
+#include <random>
+
+template<typename scalar>
+Robbie::Vector<scalar> function( const Robbie::Vector<scalar> & x )
+{
+    return x.array().pow( 2 );
+}
+
+template<typename scalar>
+void generate_training_data(
+    int n_samples, int input_size, std::vector<Robbie::Vector<scalar>> & x_train,
+    std::vector<Robbie::Vector<scalar>> & y_train )
+{
+    std::mt19937 gen                            = std::mt19937( 0 );
+    std::uniform_real_distribution<scalar> dist = std::uniform_real_distribution<scalar>( -10.0, 10.0 );
+
+    for( int i = 0; i < n_samples; i++ )
+    {
+        const auto random_lambda     = [&]( scalar x ) { return dist( gen ); };
+        Robbie::Vector<scalar> x_cur = Robbie::Vector<scalar>::Zero( input_size ).array().unaryExpr( random_lambda );
+        x_train.push_back( x_cur );
+        y_train.push_back( function( x_cur ) );
+    }
+}
 
 int main()
 {
-    fmt::print( "Hello!" );
+    using namespace Robbie;
+    std::vector<Vector<double>> x_train( 0 );
+    std::vector<Vector<double>> y_train( 0 );
+    std::vector<Vector<double>> x_test( 0 );
+    std::vector<Vector<double>> y_test( 0 );
+
+    int n_train    = 20;
+    int n_test     = n_train * 0.2;
+    int input_size = 10;
+
+    generate_training_data( n_train, input_size, x_train, y_train );
+    generate_training_data( n_test, input_size, x_test, y_test );
+
+    fmt::print( "x_train[10] = {}\n", fmt::streamed( x_train[10] ) );
+    fmt::print( "y_train[10] = {}\n", fmt::streamed( y_train[10] ) );
+
+    auto network = Network<double, LossFunctions::MeanSquareError<double>>();
+    network.add( FCLayer<double>( input_size, 100 ) );
+    network.add( ActivationLayer<double, ActivationFunctions::ReLU<double>>() );
+    network.add( FCLayer<double>( 100, 10 ) );
+    network.summary();
 
-    // Robbie::do_stuff();
+    network.fit( x_train, y_train, 25, 0.001 / 2.0, true );
 }