ray.rllib.env.env_runner_group.EnvRunnerGroup.is_policy_to_train#

EnvRunnerGroup.is_policy_to_train(policy_id: str, batch: SampleBatch | MultiAgentBatch | Dict[str, Any] | None = None) → bool[source]#

Whether given PolicyID (optionally inside some batch) is trainable.

DeveloperAPI: This API may change across minor Ray releases.