ESDistributedRollouts

class maze.train.trainers.es.distributed.es_distributed_rollouts.ESDistributedRollouts

Abstract base class of ES rollout distribution.

abstract generate_rollouts(policy: Union[maze.core.agent.policy.Policy, maze.core.agent.torch_model.TorchModel], max_steps: Optional[int], noise_stddev: float, normalization_stats: Dict[str, Tuple[numpy.ndarray, numpy.ndarray]]) → Generator[maze.train.trainers.es.distributed.es_distributed_rollouts.ESRolloutResult, None, None]

Declare a new rollout task and start producing results that can be obtained from the returned generator.

Note that different distribution strategies have different ways of balancing evaluation and training rollouts.

Parameters

• policy – Multi-step policy encapsulating the policy networks

• max_steps – Optionally limit the rollout to a number of environment steps (horizon).

• noise_stddev – The standard deviation of the applied parameter noise.

• normalization_stats – Normalization statistics as calculated by the NormalizeObservationWrapper.