public void CopyNetwork(NeuralNetworkARD _neuralNetCopy)
{
for (x = 0; x < _neuralNetCopy.axonsNet.axons.Length - 1; x++)
{
for (y = 0; y < _neuralNetCopy.axonsNet.axons[x].Length; y++)
{
for (z = 0; z < _neuralNetCopy.axonsNet.axons[x][y].Length; z++)
{
axonsNet.axons[x][y][z] = _neuralNetCopy.axonsNet.axons[x][y][z];
}
}
}
}
void Init(NeuralNetworkARD neuralNet)
{
for (i = viewerGroup.childCount - 1; i > -1; i--)
{
DestroyImmediate(viewerGroup.GetChild(i).gameObject);
}
neurons = new Image[neuralNet.layers.Length][];
neuronsValue = new Text[neuralNet.layers.Length][];
for (x = 0; x < neuralNet.layers.Length; x++)
{
neurons[x] = new Image[neuralNet.layers[x]];
neuronsValue[x] = new Text[neuralNet.layers[x]];
for (y = 0; y < neuralNet.layers[x]; y++)
{
if (neuralNet.layers[x] % 2 == 0)
{
yAdd = 1.0f;
}
else
{
yAdd = 0;
}
if (y % 2 == 0)
{
posY = y + yAdd;
}
else
{
posY = -y - 1 + yAdd;
}
neuron = Instantiate(neuronPrefab, transform.position, Quaternion.identity, viewerGroup);
neuron.anchoredPosition = new Vector2(x * decalX, posY * decalY);
neurons[x][y] = neuron.GetComponent <Image>();
neuronsValue[x][y] = neuron.transform.GetChild(0).GetComponent <Text>();
}
}
axons = new Image[neuralNet.axonsNet.axons.Length][][];
for (x = 0; x < neuralNet.axonsNet.axons.Length; x++)
{
axons[x] = new Image[neuralNet.axonsNet.axons[x].Length][];
for (y = 0; y < neuralNet.axonsNet.axons[x].Length; y++)
{
axons[x][y] = new Image[neuralNet.axonsNet.axons[x][y].Length];
for (z = 0; z < neuralNet.axonsNet.axons[x][y].Length; z++)
{
if ((neuralNet.axonsNet.axons[x].Length) % 2 == 0)
{
yAdd = 1.0f;
}
else
{
yAdd = 0;
}
if (y % 2 == 0)
{
posY = y + yAdd;
}
else
{
posY = -y - 1 + yAdd;
}
if ((neuralNet.axonsNet.axons[x][y].Length) % 2 == 0)
{
zAdd = 1.0f;
}
else
{
zAdd = 0;
}
if (z % 2 == 0)
{
posZ = z + zAdd;
}
else
{
posZ = -z - 1 + zAdd;
}
float midPosX = decalX * (x + .5f);
float midPosY = (posY + posZ) * decalY * .5f;
float zAngle = Mathf.Atan2((posY - posZ) * decalY, decalX) * Mathf.Rad2Deg;
axon = Instantiate(axonPrefab, transform.position, Quaternion.identity, viewerGroup);
axon.anchoredPosition = new Vector2(midPosX, (midPosY));
axon.eulerAngles = new Vector3(0, 0, zAngle);
axon.sizeDelta =
new Vector2(new Vector2(decalX, (posY - posZ) * decalY).magnitude * 1 - 35, 2);
axons[x][y][z] = axon.GetComponent <Image>();
}
}
}
fitness = Instantiate(fitnessPrefab, transform.position, Quaternion.identity, viewerGroup);
fitness.anchoredPosition = new Vector2(decalX * neuralNet.neurons.Length * .5f + 300, 0);
fitnessText = fitness.GetComponent <Text>();
}
