ray.rllib.callbacks.callbacks.RLlibCallback.on_episode_created#

RLlibCallback.on_episode_created(*, episode: SingleAgentEpisode | MultiAgentEpisode | EpisodeV2, worker: EnvRunner | None = None, env_runner: EnvRunner | None = None, metrics_logger: MetricsLogger | None = None, base_env: BaseEnv | None = None, env: gymnasium.Env | None = None, policies: Dict[str, Policy] | None = None, rl_module: RLModule | None = None, env_index: int, **kwargs) → None[source]#

Callback run when a new episode is created (but has not started yet!).

This method gets called after a new SingleAgentEpisode or MultiAgentEpisode instance has been created. This happens before the respective sub-environment’s reset() is called by RLlib.

  1. SingleAgentEpisode/MultiAgentEpisode created: This callback is called.

  2. Respective sub-environment (gym.Env) is reset().

  3. Callback on_episode_start is called.

  4. Stepping through sub-environment/episode commences.

Parameters:

  • episode – The newly created SingleAgentEpisode or MultiAgentEpisode. This is the episode that is about to be started with an upcoming env.reset(). Only after this reset call, the on_episode_start callback will be called.

  • env_runner – Reference to the current EnvRunner.

  • metrics_logger – The MetricsLogger object inside the env_runner. Can be used to log custom metrics after Episode creation.

  • env – The gym.Env running the episode.

  • rl_module – The RLModule used to compute actions for stepping the env. In single-agent mode, this is a simple RLModule, in multi-agent mode, this is a MultiRLModule.

  • env_index – The index of the sub-environment that is about to be reset.

  • kwargs – Forward compatibility placeholder.