private Task StartGame(EncogMaze encogMaze)
{
Task task = Task.Run(() =>
{
encogMaze.Train(TotalIterations, maxTemp, minTemp, encogCycles);
});
return(task);
}
private async void windowLessTraining_Loaded_1(object sender, RoutedEventArgs e)
{
if (isEncog)
{
EncogMaze encogMaze = new EncogMaze(maze);
encogMaze.MazeCycleComplete += Traveler_MazeCycleComplete;
//encogMaze.EncogCycleComplete += EncogMaze_EncogCycleComplete;
encogMaze.TrainingIterationComplete += EncogMaze_TrainingIterationComplete;
await StartGame(encogMaze);
}
else
{
// reset session count
//RNNMazeTraveler.ResestStaticSessionData();
traveler = new RLMMazeTraveler(maze, Learn, Temp_num_sessions, StartRandomness, EndRandomness);
traveler.MazeCycleComplete += Traveler_MazeCycleComplete;
traveler.SessionStarted += Traveler_SessionStarted;
await StartGame(traveler);
}
}
/// <summary>
/// This is where encog is being configured and trained. Call this method and pass the needed parameters to start the training.
/// </summary>
/// <param name="maze">Contains all the details of the maze that encog will be going to solve.</param>
/// <param name="gameTimeout">The given time for solving the maze per session.</param>
public static void MazeTrain(MazeInfo maze, int gameTimeout)
{
Console.WriteLine("Encog network structure:");
int defHiddenLayers = maze.Width * maze.Height; //This is the default number of hidden layer neurons if not specified.
int hiddenLayers = Util.GetInput("Number of hidden layers [default 1]: ", 1); //Getting user input for the number of hidden layers, will be set to 1(default) if not specified.
int hiddenLayersNeurons = Util.GetInput($"Hidden layers rneurons [default {defHiddenLayers}]:", defHiddenLayers); //Getting user input for the number of hidden layer neurons, will be set to "defHiddenLayers"(default) value if not specified.
var em = new EncogMaze(maze, hiddenLayers, hiddenLayersNeurons); //Create an instance of encog maze game to configure the network.
em.EncogCycleComplete += SesionComplete;
em.TrainingIterationComplete += Em_DoneTrainingIteration;
int mode = Util.GetInput("Select Encog learning method [Annealing - 0, Genetic - 1 default]: ", 1); //Gets user input for the type of encog training method, e.g. (Simulated Annealing, MLGeneticAlgorithm)
int epochs = Util.GetInput("EPOCHS to execute [default 10]: ", 10); //Gets user input for the number of epochs
int cycles = Util.GetInput((mode == 0) ? "Cycles per epoch [defualt 10]: " : "Population size [default 10]: ", 10); //Gets user input for the number of cycles
int maxTemp = 0;
int minTemp = 0;
if (mode == 0)
{
maxTemp = Util.GetInput("Max temperature [default 10]: ", 10); //Gets user input for the starting temperature
minTemp = Util.GetInput("Min temperature [default 2]: ", 2); //Gets user input for the ending temperature
}
Console.WriteLine();
System.Diagnostics.Stopwatch watch = new System.Diagnostics.Stopwatch();
watch.Start();
em.Train(mode, epochs, maxTemp, minTemp, cycles, gameTimeout); //Call the Train() method from encog game lib to start the training.
watch.Stop();
Console.WriteLine($"Elapsed: {watch.Elapsed}");
Console.ReadLine();
}
