| 117 | |
| 118 | # approximates V(s) |
| 119 | class ValueModel: |
| 120 | def __init__(self, D, hidden_layer_sizes): |
| 121 | # create the graph |
| 122 | self.layers = [] |
| 123 | M1 = D |
| 124 | for M2 in hidden_layer_sizes: |
| 125 | layer = HiddenLayer(M1, M2) |
| 126 | self.layers.append(layer) |
| 127 | M1 = M2 |
| 128 | |
| 129 | # final layer |
| 130 | layer = HiddenLayer(M1, 1, lambda x: x) |
| 131 | self.layers.append(layer) |
| 132 | |
| 133 | # inputs and targets |
| 134 | self.X = tf.placeholder(tf.float32, shape=(None, D), name='X') |
| 135 | self.Y = tf.placeholder(tf.float32, shape=(None,), name='Y') |
| 136 | |
| 137 | # calculate output and cost |
| 138 | Z = self.X |
| 139 | for layer in self.layers: |
| 140 | Z = layer.forward(Z) |
| 141 | Y_hat = tf.reshape(Z, [-1]) # the output |
| 142 | self.predict_op = Y_hat |
| 143 | |
| 144 | cost = tf.reduce_sum(tf.square(self.Y - Y_hat)) |
| 145 | # self.train_op = tf.train.AdamOptimizer(1e-2).minimize(cost) |
| 146 | # self.train_op = tf.train.MomentumOptimizer(1e-2, momentum=0.9).minimize(cost) |
| 147 | self.train_op = tf.train.GradientDescentOptimizer(1e-4).minimize(cost) |
| 148 | |
| 149 | def set_session(self, session): |
| 150 | self.session = session |
| 151 | |
| 152 | def partial_fit(self, X, Y): |
| 153 | X = np.atleast_2d(X) |
| 154 | Y = np.atleast_1d(Y) |
| 155 | self.session.run(self.train_op, feed_dict={self.X: X, self.Y: Y}) |
| 156 | |
| 157 | def predict(self, X): |
| 158 | X = np.atleast_2d(X) |
| 159 | return self.session.run(self.predict_op, feed_dict={self.X: X}) |
| 160 | |
| 161 | |
| 162 | def play_one_td(env, pmodel, vmodel, gamma): |