230503ac56636d91d53d728c247c6ea3d096eeb9,tf_agents/environments/tf_py_environment_test.py,TFPYEnvironmentTest,testOneStep,#TFPYEnvironmentTest#Any#,194

Before Change

    else:
      tf_env = tf_py_environment.TFPyEnvironment(py_env)
    _, step_data = tf_env.current_time_step()
    time_step, step_data = self.evaluate(
        tf_env.step([np.array([1])], step_data))

    self.assertAllEqual([ts.StepType.MID], time_step.step_type)
    self.assertAllEqual([0.], time_step.reward)
    self.assertAllEqual([1.0], time_step.discount)

After Change

    else:
      tf_env = tf_py_environment.TFPyEnvironment(py_env)
    time_step = tf_env.current_time_step()
    with tf.control_dependencies([time_step.step_type]):
      action = tf.constant([1])
    time_step = self.evaluate(tf_env.step(action))

    self.assertAllEqual([ts.StepType.MID], time_step.step_type)
    self.assertAllEqual([0.], time_step.reward)

Instances