GANeuralNetwork.py

import numpy as np
import os
import keras
from keras.models import Model
from keras.layers import Dense
from keras.models import Sequential, model_from_json

DATA_DIR = "./keras-nn-data/"
class GANeuralNetwork:
    def __init__(self, dimensions, weights = None, load_model = False):
        self.dimensions = dimensions

        # load model if prompted
        if load_model:
            self.load_model()
            return

        self.model = self.build_model()
        # set weights if provided
        if weights is not None:
            self.set_weights(weights)

    def shape(self):
        return np.array(self.weights).shape

    def set_weights(self, weights):
        if len(self.model.layers) != len(weights):
            print("ERROR: Weight mismatch")
            return
        for w, l in zip(weights, self.model.layers):
            l.set_weights(w)

    def weights(self):
        return [layer.get_weights() for layer in self.model.layers]

    def build_model(self):
        model = Sequential()
        model.add(
                Dense(
                    self.dimensions[1],
                    input_dim = self.dimensions[0],
                    activation = 'sigmoid',
                    kernel_initializer=keras.initializers.RandomUniform(minval=-1, maxval=1),
                    bias_initializer=keras.initializers.RandomUniform(minval=-1, maxval=1),
                )
        )
        model.add(
            Dense(
                self.dimensions[2],
                activation = 'sigmoid',
                kernel_initializer=keras.initializers.RandomUniform(minval=-1, maxval=1),
                bias_initializer=keras.initializers.RandomUniform(minval=-1, maxval=1),
            )
        )

        return model

    def load_model(self):
        with open(DATA_DIR + "ga-model.json", "r") as model_file:
            self.model = model_from_json(model_file.read())
        self.model.load_weights(DATA_DIR + "ga-model.h5")
        print("The network has been loaded from disk...")

    def save_model(self):
        # create data_dir if not yet present
        if not os.path.isdir(DATA_DIR):
            os.mkdir(DATA_DIR)

        with open(DATA_DIR + "ga-model.json", "w") as json_file:
            json_file.write(self.model.to_json())

        self.model.save_weights(DATA_DIR + "ga-model.h5")

    def run(self, X):
        return self.model.predict(X)

    def get_movement(self, trX):
        trX = np.array(trX).reshape(-1, len(trX))
        return self.run(trX)[0]