StepStateActionCriticComposer

class maze.perception.models.critics.step_state_action_critic_composer.StepStateActionCriticComposer(observation_spaces_dict: Dict[str | int, gymnasium.spaces.Dict], action_spaces_dict: Dict[str | int, gymnasium.spaces.Dict], networks: List[None | Mapping[str, Any] | Any] | Mapping[str | Type, None | Mapping[str, Any] | Any])

Each sub-step or actor gets its individual critic.

Instantiates a TorchStepStateActionCritic.

Parameters:
  • observation_spaces_dict – Dict of sub-step id to observation space.

  • action_spaces_dict – Dict of sub-step id to action space.

  • networks – Critics networks as defined in the config (either list or dictionary of object params and type).

property critic: TorchStepStateActionCritic

(overrides BaseStateActionCriticComposer)

implementation of

BaseStateActionCriticComposer