private void button_test_Click(object sender, EventArgs e)
{
chart_test_total.Series[0].Points.Clear();
listview_test.Items.Clear();
int total = int.Parse(textbox_testcount.Text), correct = 0;
AIWrapper wrapper = new AIWrapper(textbox_exepath.Text);
wrapper.SetNetwork(textbox_networkpath.Text);
wrapper.SetWeight(textbox_biasweightpath.Text);
test_res = wrapper.Predict(dataPath,
int.Parse(textbox_dataoffset.Text), dataWidth * dataWidth, total);
int i, j = 0, result, answer;
foreach (List <float> lf in test_res)
{
result = GetMaxIdx(lf);
answer = -1;
for (i = 0; i < answerSplit[j].Length; i++)
{
if (answerSplit[j][i].CompareTo(1) == 0)
{
answer = i;
break;
}
}
if (result == answer)
{
correct++;
}
//draw(panel_data_graph_2, j);
Console.WriteLine("ans : " + answer + ", res : " + result
+ "(" + (result == answer ? "O" : "X") + ")");
ListViewItem item = new ListViewItem(j++ + "");
item.SubItems.Add(answer + "");
item.SubItems.Add(result + "");
item.SubItems.Add((answer == result) ? "O" : "X");
listview_test.Items.Add(item);
//Thread.Sleep(1000);
}
chart_test_total.Series[0].Points.AddXY("Correct", correct + "");
chart_test_total.Series[0].Points.AddXY("Wrong", (total - correct) + "");
}
private void button_train_Click(object sender, EventArgs e)
{
int trainCount = int.Parse(textbox_traincount.Text);
string path = textbox_exportpath.Text;
AIWrapper wrapper = new AIWrapper(textbox_exepath.Text);
wrapper.SetNetwork(textbox_networkpath.Text);
wrapper.SetInput(textbox_datapath.Text
, int.Parse(textbox_dataoffset.Text)
, int.Parse(textbox_datalength.Text)
, dataCount);
wrapper.SetAnswer(textbox_answerpath.Text
, int.Parse(textbox_answeroffset.Text)
, int.Parse(textbox_answerlength.Text)
, answerCount);
wrapper.Train(textbox_exportpath.Text, trainCount);
}
private void button_predict_Click(object sender, EventArgs e)
{
int paint_size = panel_paint.Width;
int i, j, x, y;
//for(i = 0;i < paint_size; i++)
//{
// for(j = 0;j < paint_size; j++)
// {
// Console.Write(pixels[i][j] == 1 ? "X" : " ");
// }
// Console.WriteLine();
//}
int paintWidth = panel_paint.Width;
int scaleSize = paintWidth / dataWidth + 1;
byte[][] arr = new byte[dataWidth][];
byte[] arr2 = new byte[dataWidth * dataWidth];
for (i = 0; i < dataWidth; i++)
{
arr[i] = new byte[dataWidth];
}
for (i = 0; i < paintWidth; i++)
{
for (j = 0; j < paintWidth; j++)
{
if (pixels[i][j] == 1)
{
x = i / scaleSize;
y = j / scaleSize;
arr[i / scaleSize][j / scaleSize] = 255;
arr2[x * dataWidth + y] = 255;
}
}
}
//for(i = 0;i < arr.Length; i++)
//{
// for(j = 0;j < arr.Length; j++)
// {
// Console.Write(arr[i][j] == 255 ? "X" : "0");
// }
// Console.WriteLine();
//}
string path = Directory.GetCurrentDirectory() + @"\temp\predict.txt";
File.WriteAllBytes(path, arr2);
AIWrapper wrapper = new AIWrapper(textbox_exepath.Text);
wrapper.SetNetwork(textbox_networkpath.Text);
wrapper.SetWeight(textbox_biasweightpath.Text);
List <List <float> > res = wrapper.Predict(path, 0, dataWidth * dataWidth, 1);
int tag = 0;
foreach (List <float> lf in res)
{
foreach (float f in lf)
{
Console.WriteLine(f);
chart_predict.Series[0].Points.AddXY(tag++ + "", (f * 100) + "");
}
}
File.Delete(path);
}
