def testCompatibilityCheck(self):
r_t = tf.placeholder(tf.float32, [2])
with self.assertRaisesRegexp(
ValueError,
"PersistentQLearning: Error in rank and/or compatibility check"):
self.persistent_qlearning = rl.persistent_qlearning(
self.q_tm1, self.a_tm1, r_t, self.pcont_t, self.q_t,
self.action_gap_scale)
def testScalarCheck(self):
action_gap_scale = 2
with self.assertRaisesRegexp(
ValueError,
r"PersistentQLearning: action_gap_scale has to lie in \[0, 1\]\."):
self.persistent_qlearning = rl.persistent_qlearning(
self.q_tm1, self.a_tm1, self.r_t, self.pcont_t, self.q_t,
action_gap_scale)
def testCompatibilityCheck(self):
r_t = tf.placeholder(tf.float32, [2])
with self.assertRaisesRegexp(
ValueError,
"PersistentQLearning: Error in rank and/or compatibility check"
):
self.persistent_qlearning = rl.persistent_qlearning(
self.q_tm1, self.a_tm1, r_t, self.pcont_t, self.q_t,
self.action_gap_scale)
def testScalarCheck(self):
action_gap_scale = 2
with self.assertRaisesRegexp(
ValueError,
r"PersistentQLearning: action_gap_scale has to lie in \[0, 1\]\."
):
self.persistent_qlearning = rl.persistent_qlearning(
self.q_tm1, self.a_tm1, self.r_t, self.pcont_t, self.q_t,
action_gap_scale)
def setUp(self):
super(PersistentQLearningTest, self).setUp()
self.q_tm1 = tf.constant([[1, 2], [3, 4], [5, 6]], dtype=tf.float32)
self.a_tm1 = tf.constant([0, 1, 1], dtype=tf.int32)
self.pcont_t = tf.constant([0, 1, 0.5], dtype=tf.float32)
self.r_t = tf.constant([3, 2, 7], dtype=tf.float32)
self.q_t = tf.constant([[11, 12], [20, 16], [-8, -4]], dtype=tf.float32)
self.action_gap_scale = 0.25
self.persistent_qlearning = rl.persistent_qlearning(
self.q_tm1, self.a_tm1, self.r_t, self.pcont_t, self.q_t,
self.action_gap_scale)
def setUp(self):
super(PersistentQLearningTest, self).setUp()
self.q_tm1 = tf.constant([[1, 2], [3, 4], [5, 6]], dtype=tf.float32)
self.a_tm1 = tf.constant([0, 1, 1], dtype=tf.int32)
self.pcont_t = tf.constant([0, 1, 0.5], dtype=tf.float32)
self.r_t = tf.constant([3, 2, 7], dtype=tf.float32)
self.q_t = tf.constant([[11, 12], [20, 16], [-8, -4]],
dtype=tf.float32)
self.action_gap_scale = 0.25
self.persistent_qlearning = rl.persistent_qlearning(
self.q_tm1, self.a_tm1, self.r_t, self.pcont_t, self.q_t,
self.action_gap_scale)