ray.rllib.evaluation.worker_set.WorkerSet.sync_env_runner_states#
- WorkerSet.sync_env_runner_states(from_worker: EnvRunner | None = None, env_steps_sampled: int | None = None, timeout_s: float | None = None) None [source]#
Synchronizes the connectors of this WorkerSet’s EnvRunners.
The exact procedure works as follows: - Get all remote EnvRunners’ ConnectorV2 states. - Merge them into a resulting state. - Broadcast the resulting state back to all remote EnvRunners AND the local EnvRunner.
- Parameters:
from_worker – The EnvRunner from which to synch. If None, will try to use the local worker of this WorkerSet.