Does PPO algorithm call tianshou package directly? How to install tianshou?

gym~=0.26.2
tianshou~=0.4.10
pettingzoo
tianshou平台别安装最新版，因为它从0.5版本后就不支持gym而改为支持gymnasium了

Could you explain the program's environment? Lately, I've been encountering some errors at certain places during execution.

Thank you.

I have changed your code only the cpu to run in cpu format not skuda but it returns error: Traceback (most recent call last):
File "/home/ad/Downloads/mec_morl_multipolicy-main/train.py", line 191, in
batch = train_collector.collect(n_episode=train_num)
File "/home/ad/.local/lib/python3.10/site-packages/tianshou/data/collector.py", line 359, in collect
ptr, ep_rew, ep_len, ep_idx = self.buffer.add(
File "/home/ad/.local/lib/python3.10/site-packages/tianshou/data/buffer/manager.py", line 137, in add
ptr, ep_rew, ep_len, ep_idx = self.buffers[buffer_id]._add_index(
File "/home/ad/.local/lib/python3.10/site-packages/tianshou/data/buffer/base.py", line 209, in _add_index
if done:
File "/home/ad/.local/lib/python3.10/site-packages/tianshou/data/batch.py", line 688, in len
raise TypeError(f"Object {self} has no len()")
TypeError: Object Batch() has no len()
Have you checked and explain because I have figured out for a week

How to find the data set

您好，请问在mec_env的文件中，是需要去继承gym.env 吗，主要在训练的时候 train_envs哪里会报错，需要的是gym环境 ，而不是自定义的mec环境，但是我去继承gym后，一直bug卡在 observation——space哪里 说是NotImplementedError: Cannot convert space of type 2. Please upgrade your code to gymnasium.但是我版本已是最新的，请给予指教，万分感谢