feedforward_network.py

import numpy as np
import tensorflow as tf

def feedforward_network(inputState, inputSize, outputSize, num_fc_layers, depth_fc_layers, tf_datatype, tiled_onehots):

    #vars
    intermediate_size=depth_fc_layers
    reuse= False
    initializer = tf.contrib.layers.xavier_initializer(uniform=False, seed=None, dtype=tf_datatype)
    fc = tf.contrib.layers.fully_connected

    # make hidden layers
    for i in range(num_fc_layers):
        if(i==0):
            fc_i = fc(inputState, num_outputs=intermediate_size, activation_fn=None, 
                    weights_initializer=initializer, biases_initializer=initializer, reuse=reuse, trainable=True)
        else:
            fc_i = fc(h_i, num_outputs=intermediate_size, activation_fn=None, 
                    weights_initializer=initializer, biases_initializer=initializer, reuse=reuse, trainable=True)
        h_i = tf.nn.relu(fc_i)

    # make output layer
    z=fc(h_i, num_outputs=outputSize, activation_fn=None, weights_initializer=initializer, 
        biases_initializer=initializer, reuse=reuse, trainable=True)
    return z