private static void TestParalel()
{
Dataset dataset = new Dataset("dataset.txt");
int[] architecture = new int[] { 2, 8, 4, 3 };
NeuralNetwork nn = new NeuralNetwork(architecture);
int[] architecture2 = new int[] { 2, 8, 4, 3 };
NeuralNetwork nn2 = new NeuralNetwork(architecture2);
int[] architecture3 = new int[] { 2, 8, 4, 3 };
NeuralNetwork nn3 = new NeuralNetwork(architecture3);
GA test1 = new GA(50000, 25, 0.1, NeuralNetwork.CalculateNumOfParams(architecture), 5, -5, nn, dataset,
0.05, 0.05, 0.8, 0.12, 0.5, "test1", true);
GA test2 = new GA(50000, 25, 0.4, NeuralNetwork.CalculateNumOfParams(architecture2), 5, -5, nn2, dataset,
0.05, 0.05, 0.8, 0.12, 0.5, "test2", true);
GA test3 = new GA(50000, 15, 0.1, NeuralNetwork.CalculateNumOfParams(architecture3), 5, -5, nn3, dataset,
0.05, 0.05, 0.8, 0.12, 0.5, "test3", true);
Parallel.Invoke(
() => test1.ZapocniAlgoritam(),
() => test2.ZapocniAlgoritam(),
() => test3.ZapocniAlgoritam());
}
private static void Test3()
{
Dataset dataset = new Dataset("dataset.txt");
int[] architecture = new int[] { 2, 8, 4, 3 };
NeuralNetwork nn = new NeuralNetwork(architecture);
GA testEl2 = new GA(50000, 35, 0.05, NeuralNetwork.CalculateNumOfParams(architecture), 5, -5, nn, dataset,
0.05, 0.05, 0.8, 0.12, 0.5, "test2", true);
testEl2.ZapocniAlgoritam();
}
private static void Test5Short()
{
Dataset dataset = new Dataset("dataset.txt");
int[] architecture = new int[] { 2, 8, 3 };
NeuralNetwork nn = new NeuralNetwork(architecture);
GA testEl2 = new GA(200000, 20, 0.2, NeuralNetwork.CalculateNumOfParams(architecture), 5, -5, nn, dataset,
0.04, 0.04, 0.8, 0.12, 0.3, "prvi", false);
var param = testEl2.ZapocniAlgoritam();
nn.InspectOutput(param, dataset);
}
private static void Test5() //ovaj koristi za 2 x 8 x 4 x 3
{
Dataset dataset = new Dataset("dataset.txt");
int[] architecture = new int[] { 2, 8, 4, 3 };
NeuralNetwork nn = new NeuralNetwork(architecture);
GA testEl2 = new GA(200000, 25, 0.20, NeuralNetwork.CalculateNumOfParams(architecture), 5, -5, nn, dataset, //najbolji do sad
0.04, 0.04, 0.5, 0.22, 0.8, "prvi", false);
var param = testEl2.ZapocniAlgoritam();
nn.InspectClassification(param, dataset);
nn.WriteParamsToFile(param, "2843n1b.txt", "2843n2b.txt");
}
private static void TestLong()
{
Dataset dataset = new Dataset("dataset.txt");
int[] architecture = new int[] { 2, 8, 4, 3 };
NeuralNetwork nn = new NeuralNetwork(architecture);
GA testEl2 = new GA(800000, 25, 0.20, NeuralNetwork.CalculateNumOfParams(architecture), 5, -5, nn, dataset,
0.04, 0.04, 0.5, 0.22, 0.8, "prvi", false);
var param = testEl2.ZapocniAlgoritam();
nn.InspectClassification(param, dataset);
nn.WriteParamsToFile(param, "2843n1new.txt", "2843n2new.txt");
}
private static void Test2()
{
Dataset dataset = new Dataset("dataset.txt");
int[] architecture = new int[] { 2, 8, 4, 3 };
NeuralNetwork nn = new NeuralNetwork(architecture);
//GA testEl2 = new GA(50000, 60, 0.1, NeuralNetwork.CalculateNumOfParams(architecture), 3, -3, nn, dataset,
// 0.05, 0.05, 0.6, 1, 1, "sample string"); //first try
GA testEl2 = new GA(50000, 20, 0.1, NeuralNetwork.CalculateNumOfParams(architecture), 5, -5, nn, dataset,
0.05, 0.05, 0.8, 0.12, 0.5, "prvi", true);
testEl2.ZapocniAlgoritam();
}