To solve non-linear problems, we need to connect several perceptrons together to create a neural network. In the example, the multilayer perceptron resolve a XOR operation.
The inputs will be propagated between several layers of perceptrons and output a value as a result.
See also my projects :
Create a new neural network with his number of inputs, hidden layers and outputs.
var neural_network = NeuralNetwork.new(INPUTS, HIDDEN_LAYERS, OUPUTS)
To train the neural network, pass the input values and the expected result to the training function.
neural_network.train([INPUT_1, INPUT_X...], EXEPTED_RESULT)
This project require Godot Engine 3.3.
- The source code is available under the MIT license.