BaseModelComposer

class maze.perception.models.model_composer.BaseModelComposer(action_spaces_dict: Dict[str | int, gymnasium.spaces.Dict], observation_spaces_dict: Dict[str | int, gymnasium.spaces.Dict], agent_counts_dict: Dict[str | int, int], distribution_mapper_config: None | Mapping[str, Any] | Any)

Abstract baseclass and interface definitions for model composers.

Model composers encapsulate the set of policy and critic networks along with the distribution mapper.

Parameters:

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

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

• distribution_mapper_config – Distribution mapper configuration.

abstract classmethod check_model_config(model_config: None | Mapping[str, Any] | Any) → None

Asserts the provided model config for consistency. :param model_config: The model config to check.

abstract property critic: TorchStateCritic | None

The critic model.

property distribution_mapper: DistributionMapper

The DistributionMapper, mapping the action heads to distributions.

abstract property policy: TorchPolicy | None

Policy networks.

save_models() → None

Save the policies and critics as pdfs.