protected void RunStep()
{
trainer.Step(environment);
bool reset = trainer.Record(environment);
episodePoint += environment.LastReward();
//reset if end
if (reset && training)
{
environment.Reset();
episodesThisTrain++;
episodePointAve.AddValue(episodePoint);
if (episodePointAve.JustUpdated)
{
scoreUI.text = "(PPO, Continuous) Average Reward:" + episodePointAve.Average;
}
episodePoint = 0;
if (episodesThisTrain >= episodeToRunForEachTrain)
{
trainer.TrainAllData(minibatch, iterationForEachTrain);
//record and print the loss
print("Training Loss:" + trainer.LastLoss);
trainedCount++;
trainer.ClearData();
episodesThisTrain = 0;
}
}
}
protected void RunStep()
{
trainer.Step(environment);
bool reset = trainer.Record(environment);
episodePoint += environment.LastReward();
//reset if end
if (reset && training)
{
currentEpisode++;
if (environment.GameWinPlayer == 0)
{
leftWin++;
winningRate50Left.AddValue(1);
}
else
{
rightWin++;
winningRate50Left.AddValue(0);
}
environment.Reset();
episodesThisTrain++;
episodePointAve.AddValue(episodePoint);
episodePoint = 0;
if (episodesThisTrain >= episodeToRunForEachTrain)
{
trainer.TrainAllData(minibatch, iterationForEachTrain);
//record and print the loss
print("Training Loss:" + trainer.LastLoss);
trainedCount++;
trainer.ClearData();
episodesThisTrain = 0;
}
}
}