d9ef1ee4303a0862a3a127bfea954f06811b06ba,tf_agents/policies/q_policy_test.py,QPolicyTest,testUpdate,#QPolicyTest#,156

Before Change

    self.assertAllEqual([[1]], self.evaluate(mode))

  def testUpdate(self):
    tf.compat.v1.set_random_seed(1)
    policy = q_policy.QPolicy(
        self._time_step_spec, self._action_spec, q_network=DummyNet())
    new_policy = q_policy.QPolicy(
        self._time_step_spec, self._action_spec, q_network=DummyNet())

After Change

    action = self.evaluate(action_step.action)
    new_action = self.evaluate(new_action_step.action)
    self.assertTrue(np.all(action >= 0) and np.all(action <= 1))
    self.assertTrue(np.all(new_action >= 0) and np.all(new_action <= 1))
    self.assertAllEqual(action, new_action)

  def testActionSpecsCompatible(self):

Instances