ACLocalRunner

class maze.train.trainers.common.actor_critic.actor_critic_runners.ACLocalRunner(state_dict_dump_file: str, dump_interval: Optional[int], spaces_config_dump_file: str, normalization_samples: int, trainer_class: Union[str, type], concurrency: int, eval_concurrency: int)

Runner for locally distributed training, based on SubprocVectorEnv.

create_distributed_env(env_factory: Callable[], Union[maze.core.env.maze_env.MazeEnv, maze.core.env.structured_env.StructuredEnv]], concurrency: int, logging_prefix: str) → maze.train.parallelization.vector_env.subproc_vector_env.SubprocVectorEnv

create multi-process env distribution