使用openai gym库的API mix问题

在使用训练好的模型对gym中的CartPole-v1时，写了下面的代码

obs = env.reset()
action,_ = model.predict(obs)

报了个大错，贼长，如下：

ValueError                                Traceback (most recent call last)
Cell In[43], line 1
----> 1 action,_ = model.predict(obs)

File c:\users\dell\appdata\local\programs\python\python39\lib\site-packages\stable_baselines3\common\base_class.py:556, in BaseAlgorithm.predict(self, observation, state, episode_start, deterministic)
    536 def predict(
    537     self,
    538     observation: Union[np.ndarray, Dict[str, np.ndarray]],
   (...)
    541     deterministic: bool = False,
    542 ) -> Tuple[np.ndarray, Optional[Tuple[np.ndarray, ...]]]:
    543     """
    544     Get the policy action from an observation (and optional hidden state).
    545     Includes sugar-coating to handle different observations (e.g. normalizing images).
   (...)
    554         (used in recurrent policies)
    555     """
--> 556     return self.policy.predict(observation, state, episode_start, deterministic)

File c:\users\dell\appdata\local\programs\python\python39\lib\site-packages\stable_baselines3\common\policies.py:357, in BasePolicy.predict(self, observation, state, episode_start, deterministic)
    354 # Check for common mistake that the user does not mix Gym/VecEnv API
    355 # Tuple obs are not supported by SB3, so we can safely do that check
    356 if isinstance(observation, tuple) and len(observation) == 2 and isinstance(observation[1], dict):
--> 357     raise ValueError(
    358         "You have passed a tuple to the predict() function instead of a Numpy array or a Dict. "
    359         "You are probably mixing Gym API with SB3 VecEnv API: `obs, info = env.reset()` (Gym) "
    360         "vs `obs = vec_env.reset()` (SB3 VecEnv). "
    361         "See related issue https://github.com/DLR-RM/stable-baselines3/issues/1694 "
    362         "and documentation for more information: https://stable-baselines3.readthedocs.io/en/master/guide/vec_envs.html#vecenv-api-vs-gym-api"
    363     )
    365 obs_tensor, vectorized_env = self.obs_to_tensor(observation)
    367 with th.no_grad():

ValueError: You have passed a tuple to the predict() function instead of a Numpy array or a Dict. You are probably mixing Gym API with SB3 VecEnv API: `obs, info = env.reset()` (Gym) vs `obs = vec_env.reset()` (SB3 VecEnv). See related issue https://github.com/DLR-RM/stable-baselines3/issues/1694 and documentation for more information: https://stable-baselines3.readthedocs.io/en/master/guide/vec_envs.html#vecenv-api-vs-gym-api

总结一下就是两个API混着用，需要输入列表结果输入的是元组，所以报错了。
找遍了全网没找到怎么改，问了gpt也没用。
改了好久突然发现obs, info = env.reset()这句代码，于是去看了下源文件中的定义（https://www.gymlibrary.dev/api/core/#gym.Env.reset）

observation (object)
 – Observation of the initial state. This will be an element of observation_space (typically a numpy array) and is analogous to the observation returned by step().

info (dictionary)
 – This dictionary contains auxiliary information complementing observation. It should be analogous to the info returned by step().

返回的还真不是列表，是由一个numpy array(obs)和一个dictionary(info)组成的元组。
这就好办了，用两个变量分别接收就好了，修改如下

obs,info = env.reset()
action, _ = model.predict(obs)

用type函数看看数据类型

诶对了，然后再放进去跑就行了

更新2024.5.10--------------------------------------------------------------------------------------
原来是之前

env = gym.make(environment_name, render_mode="human")
env = DummyVecEnv([lambda:env])
model = PPO('MlpPolicy', env, verbose = 1,tensorboard_log=log_path)

第二行代码没加，导致冲突了，加了第二行代码可以不用上述方法