Happy

v4 version can be `make` directly after [`pip install gym[mujoco]`](https://github.com/openai/gym#mujoco-environments) without compiling.

[`num_updates`](https://github.com/PaddlePaddle/PARL/blob/f75edc407173cf34f8855e453606bc98f151b8e8/examples/PPO/train.py#L56) 仅与训练步数相关（仅用在定义[learn的次数](https://github.com/PaddlePaddle/PARL/blob/f75edc407173cf34f8855e453606bc98f151b8e8/examples/PPO/train.py#L90)及对应的[lr更新](https://github.com/PaddlePaddle/PARL/blob/f75edc407173cf34f8855e453606bc98f151b8e8/examples/PPO/agent.py#L35))，需要根据训练的[总步数](https://github.com/PaddlePaddle/PARL/blob/f75edc407173cf34f8855e453606bc98f151b8e8/examples/PPO/atari_config.py#L26)以及[batch_size](https://github.com/PaddlePaddle/PARL/blob/f75edc407173cf34f8855e453606bc98f151b8e8/examples/PPO/train.py#L55)计算，严格意义不属于需要自主定义的参数。

Thanks, but there is No need for so much `print` and too much useless information

Fixing in PR https://github.com/PaddlePaddle/PARL/pull/969

Fixed in PR https://github.com/PaddlePaddle/PARL/pull/942

Refer to issue【 https://github.com/PaddlePaddle/PARL/issues/817 第四课PG作业Pong报错：WinError 126找不到指定模块】 or 【 https://github.com/PaddlePaddle/PARL/issues/895 最近更新的第四课作业的动态图，环境配置问题】

Sorry for the delay. Haven't tried that. You'll have to change the [world env](https://github.com/ShuaibinLi/RL_CARLA/blob/350fc0c3d0ff8398d7fb6f0774c0e79ab54ebef5/gym_carla/gym_carla/envs/carla_env.py#LL140C48-L140C48) to add more info alongside ego vehicle. That'd be cool if you gave it a shot.

Thanks for your torch-based training. Can you provide the results together in another branch (e.g. torch-based)?

You have to find the car on the map according to the coordinates in [gym_carla/gym_carla/envs/coordinates.py](https://github.com/ShuaibinLi/RL_CARLA/blob/main/gym_carla/gym_carla/envs/coordinates.py).

+1