def testVariableBatchSize(self):
q_tm1 = tf.placeholder(tf.float32, shape=[None, 3])
a_tm1 = tf.placeholder(tf.int32, shape=[None])
pcont_t = tf.placeholder(tf.float32, shape=[None])
r_t = tf.placeholder(tf.float32, shape=[None])
v_t = tf.placeholder(tf.float32, shape=[None])
loss_op, extra_ops = rl.qv_learning(q_tm1, a_tm1, r_t, pcont_t, v_t)
# Check static shapes.
self.assertEqual(loss_op.get_shape().as_list(), [None])
self.assertEqual(extra_ops.td_error.get_shape().as_list(), [None])
self.assertEqual(extra_ops.target.get_shape().as_list(), [None])
# Check runtime shapes.
batch_size = 11
feed_dict = {
q_tm1: np.random.random([batch_size, 3]),
a_tm1: np.random.randint(0, 3, [batch_size]),
pcont_t: np.random.random([batch_size]),
r_t: np.random.random(batch_size),
v_t: np.random.random(batch_size),
}
with self.test_session() as sess:
loss, td_error, target = sess.run(
[loss_op, extra_ops.td_error, extra_ops.target],
feed_dict=feed_dict)
self.assertEqual(loss.shape, (batch_size, ))
self.assertEqual(td_error.shape, (batch_size, ))
self.assertEqual(target.shape, (batch_size, ))
def setUp(self):
super(QVTest, self).setUp()
self.q_tm1 = tf.constant([[1, 1, 0], [1, 1, 0]], dtype=tf.float32)
self.a_tm1 = tf.constant([0, 1], dtype=tf.int32)
self.pcont_t = tf.constant([0, 1], dtype=tf.float32)
self.r_t = tf.constant([1, 1], dtype=tf.float32)
self.v_t = tf.constant([1, 3], dtype=tf.float32)
self.loss_op, self.extra_ops = rl.qv_learning(self.q_tm1, self.a_tm1,
self.r_t, self.pcont_t,
self.v_t)
def testRankCheck(self):
q_tm1 = tf.placeholder(tf.float32, [None])
with self.assertRaisesRegexp(
ValueError,
"QVLearning: Error in rank and/or compatibility check"):
rl.qv_learning(q_tm1, self.a_tm1, self.r_t, self.pcont_t, self.v_t)