public TrainingParameters()
{
InitializeComponent();
Data = new TrainingRate();
DataContext = Data;
}
public void PrepareTrainingData(TrainingRate learningParameters, int epoch, ref int index, ref bool stop)
{
if (learningParameters.Distorted)
{
if (((epoch-1) % learningParameters.SameDistortionsForNEpochs) == 0)
{
bool parameterChanged = false;
if (SeverityFactor != learningParameters.SeverityFactor)
{
SeverityFactor = learningParameters.SeverityFactor;
parameterChanged = true;
}
if (MaxScaling != learningParameters.MaxScaling)
{
MaxScaling = learningParameters.MaxScaling;
parameterChanged = true;
}
if (MaxRotation != learningParameters.MaxRotation)
{
MaxRotation = learningParameters.MaxRotation;
parameterChanged = true;
}
if (ElasticSigma != learningParameters.ElasticSigma)
{
ElasticSigma = learningParameters.ElasticSigma;
parameterChanged = true;
}
if (ElasticScaling != learningParameters.ElasticScaling)
{
ElasticScaling = learningParameters.ElasticScaling;
parameterChanged = true;
}
if (parameterChanged)
CalculateGaussianKernel();
PreparedTrainingPatterns = new DoubleImageData[TrainingPatternsCount];
for (index = 0; index < TrainingPatternsCount; index++)
{
if (stop)
break;
PreparedTrainingPatterns[index] = new DoubleImageData(TrainingPatterns[index].Label, GetDistortedPattern(index, true));
}
}
}
else
{
PreparedTrainingPatterns = TrainingPatterns;
}
}
public void AddGlobalTrainingRate(TrainingRate rate, Boolean clear = true)
{
if (TrainingRates == null)
TrainingRates = new List<TrainingRate>();
if ((TrainingRates != null) && (clear))
TrainingRates.Clear();
if (rate.DecayFactor == 1D)
{
TrainingRates.Add(rate);
int numEpochs = 0;
foreach (TrainingRate aRate in TrainingRates)
{
numEpochs += aRate.Epochs;
}
}
else
{
// Decaying Training Rate
if (rate.Epochs < rate.DecayAfterEpochs)
rate.DecayAfterEpochs = rate.Epochs;
int totIteration = rate.Epochs / rate.DecayAfterEpochs;
double newRating = rate.Rate;
for (int i = 0; i < totIteration; i++)
{
TrainingRates.Add(new TrainingRate(newRating, rate.DecayAfterEpochs, rate.MinimumRate, rate.DecayFactor, rate.DecayAfterEpochs, rate.WeightSaveTreshold, rate.Distorted, rate.SameDistortionsForNEpochs, rate.SeverityFactor, rate.MaxScaling, rate.MaxRotation, rate.ElasticSigma, rate.ElasticScaling));
if (newRating * rate.DecayFactor > rate.MinimumRate)
newRating *= rate.DecayFactor;
else
newRating = rate.MinimumRate;
}
if ((totIteration * rate.DecayAfterEpochs) < rate.Epochs)
{
TrainingRates.Add(new TrainingRate(newRating, rate.Epochs - (totIteration * rate.DecayAfterEpochs), rate.MinimumRate, rate.DecayFactor, rate.DecayAfterEpochs, rate.WeightSaveTreshold, rate.Distorted, rate.SameDistortionsForNEpochs, rate.SeverityFactor, rate.MaxScaling, rate.MaxRotation, rate.ElasticSigma, rate.ElasticScaling));
}
}
}
public TrainingParametersEditor()
{
Data = new TrainingRate(true);
InitializeComponent();
}
private void TestingTask()
{
avgMSE = 0;
misrecognitions = 0;
curIndex = 0;
testingPageViewModel.CurrentState = TestingStates.PreparingData;
TrainingRate rate = new TrainingRate(testingPageViewModel.UseDistortions);
if (testingPageViewModel.UseDistortions)
{
rate.SeverityFactor = testingPageViewModel.SeverityFactor;
rate.MaxScaling = testingPageViewModel.MaxScaling;
rate.MaxRotation = testingPageViewModel.MaxRotation;
rate.ElasticScaling = testingPageViewModel.ElasticScaling;
rate.ElasticSigma = testingPageViewModel.ElasticSigma;
}
pageViewModel.DataProvider.PrepareTrainingData(rate, 1, ref curIndex, ref testingStopped);
testingPageViewModel.CurrentState = TestingStates.Testing;
TestPattern(ref curIndex, ref testingStopped);
testingPageViewModel.CurrentState = TestingStates.Idle;
while (testingPageViewModel.CurrentState == TestingStates.Idle)
{
Thread.SpinWait(1);
if (testingStopped)
break;
}
testingStopped = true;
}
