How easy would it be to give different observations to the actor and the ciritc?

[Jan-Loettgen]

Hi all,

I am currently trying to run some experiments, that requires feeding different observations into the PPO actor and critic. Previsouly I have been using Stable-Baselines 3, but it seems that it would be diffficult to change the code such that this would be possible. I am wondering whether something like this would be easier in SKRL?

Any tips would be appreciated, thanks