quantumiracle/Popular-RL-Algorithms

Does sac_v2_lstm support Pendulum-v0?

zhaoguangyuan123 opened this issue · 1 comments

I am trying to run sac_v2_lstm.py with python sac_v2_lstm.py --train.

I change the 'ENV = ['Reacher', 'Pendulum-v0', 'HalfCheetah-v2'][2]
' in original code to 'ENV = ['Reacher', 'Pendulum-v0', 'HalfCheetah-v2'][1]'.

The error i got

action | [0.20995763] Traceback (most recent call last): File "sac_v2_lstm.py", line 267, in <module> next_state, reward, done, _ = env.step(action) File "/usr/local/anaconda3/lib/python3.8/site-packages/gym/core.py", line 292, in step return self.env.step(self.action(action)) File "/usr/local/anaconda3/lib/python3.8/site-packages/gym/core.py", line 295, in action raise NotImplementedError NotImplementedError

Is it because for the POMDP we need particularly gym env?

Hi,

I think it's not the reason you mentioned.
I notice that you're using Python 3.8 but I run the code with Python 3.6.8 without such errors. But it's most likely to be the gym version problem, and I'm using gym==0.10.9. Take a look at the requirements.txt. It's basically a replicate of my local environment.
You can try:

pip uninstall gym
pip install gym==0.10.9