ActionRecordRolloutRunner

class maze.core.rollout.action_record_rollout_runner.ActionRecordRolloutRunner(max_episode_steps: int, deterministic: bool, action_record_path: str, normalization_samples: int, n_processes: int, verbose: bool)

Parallel rollout runner that allows to collect features by replaying pre-computed action records.

Parameters

• max_episode_steps – The maximum number of agent actions to take (careful these are not internal env steps).

• deterministic – Deterministic or stochastic action sampling.

• action_record_path – Path to action records.

• normalization_samples – Number of samples (=steps) to collect normalization statistics.

• n_processes – Count of processes to spread the rollout across.

• verbose – If True debug messages are printed to the command line.

run_with(env: Union[None, Mapping[str, Any], Any], wrappers: Union[List[Union[None, Mapping[str, Any], Any]], Mapping[Union[str, Type], Union[None, Mapping[str, Any], Any]]], agent: Union[None, Mapping[str, Any], Any])

(overrides ParallelRolloutRunner)

Run the parallel rollout in multiple worker processes.