Denys88 / rl_games

RL implementations
MIT License
800 stars 135 forks source link

question about central value #279

Open JoseBarreiros-TRI opened 3 months ago

JoseBarreiros-TRI commented 3 months ago

@Denys88 thank you for the nice repo. I noticed you use a central value network when using asymmetric actor-critic. Could you please elaborate on what the central value net is exactly doing? Is this just the critic net?

Denys88 commented 3 months ago

yes it is a critic network but with additional inputs. For example policy can have only a few parameters as obs but critic can have the whole world state because we don't use critic during inference.