DataLoadWorker¶
-
class
maze.train.trainers.imitation.parallel_loaded_im_data_set.DataLoadWorker¶ Data loading worker used to map states to actual observations.
-
static
run(env_factory: Callable, trajectory_file_paths: List[Union[pathlib.Path, str]], reporting_queue: multiprocessing.context.BaseContext.Queue) → None¶ Load trajectory data from the provided trajectory file paths. Report exceptions to the main process.
- Parameters
env_factory – Function for creating an environment for MazeState and MazeAction conversion.
trajectory_file_paths – Which trajectory data files should this worker load and process.
reporting_queue – Queue for reporting loaded data and exceptions back to the main process.
-
static