ENV_STATE for QMIX

I have seen some example on how to run QMIX on RLLIB. For instance Playing the QMIX Two-step game on Ray – Bytefreaks.net or chappers. However, I have not found any implementation of QMIX for continuous observation spaces. Is that possible? In my case, I have the following observation spaces:

self.observation_space_DA = spaces.Box(low=-np.inf, high=np.inf, shape=(212,),dtype=np.float32)
self.observation_space_BM = spaces.Box(low=-np.inf, high=np.inf, shape=(102,),dtype=np.float32)

How can I define ENV_STATE in this case?

Hi, I don’t know so much about QMIX, but in the documentation you can see that the implementation of QMIX in, at least, RLlib don’t works for continuous action spaces.