示例#1
0
        public static Image <Rgba32> GenerateImage(List <List <Node> > network, Dictionary <string, bool> state, int size)
        {
            var img2 = new Image <Rgba32>(size, size);

            for (int x = 0; x < size; x++)
            {
                for (int y = 0; y < size; y++)
                {
                    double outx = x / (double)size * 2.0 - 1.0;
                    double outy = y / (double)size * 2.0 - 1.0;

                    var input    = MLHelper.ConstructInput(state, outx, outy);
                    var lastNode = NetworkBuilder.ForwardProp(network, input);

                    //byte r = (byte)(255 * (1 - lastNode));
                    //byte g = 0;
                    //byte b = 0;

                    //=MAX((1-((A2+1)/2) / 2 * 3) *255; 0)
                    byte r = (byte)Math.Max((1 - ((lastNode + 1) / 2.0) / 2.0 * 3.0) * 255.0, 0);
                    //=MAX((1-((A2+1)/2) / 2 * 6) *128; 0)
                    byte g = (byte)Math.Max((1 - ((lastNode + 1) / 2.0) / 2.0 * 6.0) * 128.0, 0);
                    //=MIN(MAX((((A2+1 - 0,6666)/2 / 2 *3) ) *255; 0); 255)
                    byte b = (byte)Math.Min(Math.Max((((lastNode + 1 - (2.0 / 3.0)) / 2.0 / 2.0 * 3.0)) * 255.0, 0), 255);

                    img2[x, y] = new Rgba32(r, g, b);

                    //var pntX = ((pnt.X + 1.0) / 2.0) * size;
                    //var pntY = ((pnt.Y + 1.0) / 2.0) * size;
                }
            }

            return(img2);
        }
示例#2
0
        public static double GetLoss(List <List <Node> > network, Dictionary <string, bool> state, List <Puntje> dataPoints)
        {
            double loss = 0;

            for (var i = 0; i < dataPoints.Count; i++)
            {
                var dataPoint = dataPoints[i];
                var input     = MLHelper.ConstructInput(state, dataPoint.X, dataPoint.Y);
                var output    = NetworkBuilder.ForwardProp(network, input);
                loss += Errors.SQUARE.Error(output, dataPoint.Label);
            }
            return(loss / dataPoints.Count);
        }