mtenv.envs package

Submodules

mtenv.envs.registration module

class mtenv.envs.registration.MultiEnvRegistry[source]

Bases: gym.envs.registration.EnvRegistry

register(id: str, **kwargs: Any) → None[source]
class mtenv.envs.registration.MultitaskEnvSpec(id: str, entry_point: Optional[str] = None, reward_threshold: Optional[int] = None, kwargs: Optional[Dict[str, Any]] = None, nondeterministic: bool = False, max_episode_steps: Optional[int] = None, test_kwargs: Optional[Dict[str, Any]] = None)[source]

Bases: gym.envs.registration.EnvSpec

A specification for a particular instance of the environment. Used to register the parameters for official evaluations.

Parameters
  • id (str) – The official environment ID

  • entry_point (Optional[str]) – The Python entrypoint of the environment class (e.g. module.name:Class)

  • reward_threshold (Optional[int]) – The reward threshold before the task is considered solved

  • kwargs (dict) – The kwargs to pass to the environment class

  • nondeterministic (bool) – Whether this environment is non-deterministic even after seeding

  • max_episode_steps (Optional[int]) – The maximum number of steps that an episode can consist of

  • test_kwargs (Optional[Dict[str, Any]], optional) – Dictionary to specify parameters for automated testing. Defaults to None.

property kwargs
mtenv.envs.registration.make(id: str, **kwargs: Any) → gym.core.Env[source]
mtenv.envs.registration.register(id: str, **kwargs: Any) → None[source]
mtenv.envs.registration.spec(id: str)mtenv.envs.registration.MultitaskEnvSpec[source]

Module contents