Ejemplo n.º 1
0
        'scene_name': 'bedroom_05',
        'terminal_state_id': int(task_scope),
        #'initial_state': EVAL_INIT_LOC,
      })
      real_target_xz.append([xz_numpy[int(task_scope)][0], xz_numpy[int(task_scope)][1]])
      ep_rewards = []
      ep_lengths = []
      ep_collisions = []

      scopes = [network_scope, scene_scope, task_scope]
      #time.sleep(5)
      if 1:
          time.sleep(1)
          cv2.imshow('target image', env.observation_target)
          cv2.waitKey(0)
      viewer = SimpleImageViewer()
      viewer.imshow(env.observation,str(0))
      time.sleep(5)
      for i_episode in range(NUM_EVAL_EPISODES):

        env.reset()
        current_idindex=env.current_state_id
        terminal = False
        ep_reward = 0
        ep_collision = 0
        ep_t = 0
        ep_action = []

        show_target = []
        max_value = []
        max_index = 0
Ejemplo n.º 2
0
                        random_start=False,
                        random_terminal=False,
                        whe_show=False,
                        terminal_id=20,
                        start_id=0,
                        whe_use_image=True,
                        whe_flatten=False,
                        num_of_frames=1)

    # manually disable terminal states
    # env.terminals = np.zeros_like(env.terminals)
    # env.terminal_states, = np.where(env.terminals)
    # env.reset()

    # 命令参数初始化
    human_agent_action = None
    human_wants_restart = False
    stop_requested = False

    viewer = SimpleImageViewer()
    viewer.imshow(env.observation)
    viewer.window.on_key_press = key_press

    print("Use arrow keys to move the agent.")
    print("Press R to reset agent\'s location.")
    print("Press Q to quit.")

    rollout(env)

    print("Goodbye.")
Ejemplo n.º 3
0
            #fo.write('%s \n', %task_scope)

            env = Environment({
                'scene_name': scene_scope,
                'terminal_state_id': task_scope[0],
                'checkpoint_state_id': task_scope[1]
            })
            ep_rewards = []
            ep_lengths = []
            ep_collisions = []

            scopes = [network_scope, scene_scope, task_scope]

            print('evaluation: %s %s' % (scene_scope, task_scope))

            viewer = SimpleImageViewer()
            #NUM_EVAL_EPISODES
            for i_episode in range(2):

                env.reset()
                terminal = False
                ep_reward = 0
                ep_collision = 0
                ep_t = 0

                f.write(str(counter * 5 + i_episode) + ': [')
                path_x = []
                path_y = []
                path_x.append(int(env.x * 2))
                path_y.append(int(env.z * 2))
                int(env.z * 2)
Ejemplo n.º 4
0
if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument("-s",
                        "--scene_dump",
                        type=str,
                        default="./data/nnew1.h5",
                        help="path to a hdf5 scene dump file")
    parser.add_argument("-n",
                        "--n_episode",
                        type=int,
                        default=20,
                        help="number of episode to run")
    args = parser.parse_args()

    print("Loading scene dump {}".format(args.scene_dump))
    env = THORDiscreteEnvironment({'h5_file_path': args.scene_dump})

    viewer = SimpleImageViewer()

    results = []
    for i in range(args.n_episode):
        env.reset()
        graph = build_graph()
        results.append(navigate(env, graph))

    print("Success for %s times out of %s episode" %
          (len(np.array(results).nonzero()), args.n_episode))

    viewer.close()
Ejemplo n.º 5
0
    print("Loading scene dump {}".format(args.scene_dump))
    env = THORDiscreteEnvironment({
        'h5_file_path': args.scene_dump,
        'initial_state': args.start,
    })

    # manually disable terminal states
    env.terminals = np.zeros_like(env.terminals)
    env.terminal_states, = np.where(env.terminals)
    env.reset()

    human_agent_action = None
    human_wants_restart = False
    stop_requested = False
    action_list = [int(i)
                   for i in args.action.split(',')] if args.action else None
    action_idx = 0
    save_img = None

    viewer = SimpleImageViewer(save_dir=args.save_dir)
    viewer.imshow(env.observation, save_img=0)
    viewer.window.on_key_press = key_press

    print("Use arrow keys to move the agent.")
    print("Press R to reset agent\'s location.")
    print("Press Q to quit.")

    rollout(env)

    print("Goodbye.")
Ejemplo n.º 6
0
  print("Loading scene dump {}".format(args.scene_dump))
  env = THORDiscreteEnvironment({
    'h5_file_path': args.scene_dump,
    'initial_state': args.start,
  })

  # manually disable terminal states
  env.terminals = np.zeros_like(env.terminals)
  env.terminal_states, = np.where(env.terminals)
  env.reset()

  human_agent_action = None
  human_wants_restart = False
  stop_requested = False
  action_list = [int(i) for i in args.action.split(',')] if args.action else None
  action_idx = 0
  save_img = None

  viewer = SimpleImageViewer(save_dir=args.save_dir)
  viewer.imshow(env.observation, save_img=0)
  viewer.window.on_key_press = key_press

  print("Use arrow keys to move the agent.")
  print("Press R to reset agent\'s location.")
  print("Press Q to quit.")

  rollout(env)

  print("Goodbye.")
Ejemplo n.º 7
0
                'terminal_state_id': int(task_scope),
                #'initial_state': EVAL_INIT_LOC,
            })
            real_target_xz.append(
                [xz_numpy[int(task_scope)][0], xz_numpy[int(task_scope)][1]])
            ep_rewards = []
            ep_lengths = []
            ep_collisions = []

            scopes = [network_scope, scene_scope, task_scope]
            #time.sleep(5)
            if 1:
                time.sleep(1)
                cv2.imshow('target image', env.observation_target)
                cv2.waitKey(0)
            viewer = SimpleImageViewer()
            viewer.imshow(env.observation, str(0))
            time.sleep(5)
            for i_episode in range(NUM_EVAL_EPISODES):

                env.reset()
                current_idindex = env.current_state_id
                terminal = False
                ep_reward = 0
                ep_collision = 0
                ep_t = 0
                ep_action = []

                show_target = []
                max_value = []
                max_index = 0