public CharRecognSpecies(CookedSamples cooked, int netInputs, params int[] neuronCounts)
{
charNetwork = new CharRecognizerNetwork();
charNetwork.AllocateNetwork(false, netInputs, neuronCounts);
charNetwork.SetWeights(WeightsFromChromosomes(), );
charNetwork = net;
samples = cooked;
samplesToPass = cooked.Length;
}
private void tbnTrainBackProp_Click(object sender, EventArgs e)
{
if (backgroundWorkerTrain.IsBusy)
{
tbnTrainBackProp.Text = "Cancelling...";
backgroundWorkerTrain.CancelAsync();
return;
}
double.TryParse(txtA.Text, out charNet.const_a);
txtA.Text = charNet.const_a.ToString();
double.TryParse(txtBias.Text, out charNet.bias);
txtBias.Text = charNet.bias.ToString();
double.TryParse(txtEta.Text, out charNet.const_Eta);
txtEta.Text = charNet.const_Eta.ToString();
progressBar1.Maximum = charNet.MaxEpoch;
progressBar1.Value = 0;
int inputs = nx * ny;
int outputs = chars.Length;
int[] NeuronCounts = { inputs, (inputs + outputs) / 2, outputs };
charNet.AllocateNetwork(inputs, NeuronCounts); // inputs may not match the number of neurons in the 1st layer
charNet.FillSmallRandom(5e-2);
CookSamples(out xCookedSamples, out yCookedSamples);
// Train();
tbnTrainBackProp.Text = "Cancel";
backgroundWorkerTrain.RunWorkerAsync();
}