Skip to content

GymVectorEnv

Full path: schola.gym.env.GymVectorEnv

A Gym Vector Environment that wraps a Schola Environment.

GymVectorEnv(simulator, protocol, verbosity = 0,
             autoreset_mode = <AutoresetMode SAME_STEP>)

Parameters

  • simulator (BaseSimulator) - Simulator managing the Unreal (or other) process lifecycle.

  • protocol (BaseRLProtocol) - Protocol used to talk to the running simulation.

  • verbosity (int, default: 0) - Reserved verbosity level for future logging hooks.

  • autoreset_mode (str, default: AutoresetMode.SAME_STEP) - Passed to the protocol as auto-reset behavior when episodes end.

Methods

init

__init__(simulator, protocol, verbosity = 0,
         autoreset_mode = <AutoresetMode SAME_STEP>)

Parameters

  • simulator (BaseSimulator)

  • protocol (BaseRLProtocol)

  • verbosity (int)

  • autoreset_mode (str)

close

close(**kwargs)

get_attr

get_attr(name)

Get an attribute from the environment.

Parameters

  • name (str) - The name of the attribute to get.

reset

reset(seed = None, options = None)

Parameters

  • seed (None)

  • options (Dict)

step

step(actions)

Parameters

  • actions (Dict)

unbatch_actions

unbatch_actions(actions)

Unbatch actions from Dict[ObsID,Batched] to a nested Dict[EnvId,Dict[AgentId,Dict[ObsId,Value]]], effectively moving the env, and agent dimensions into Dictionaries.

Parameters

  • actions (Dict) - The batched actions to unbatch.

Attributes

reset_infos

reset_infos : List[Dict[str, Any]]

One info dict per flattened agent slot (length num_envs), updated on each reset(). Values are strings from the simulator in typical setups. For Gymnasium’s batched layout (arrays plus _<key> masks), use the infos value returned from reset(); that dict is built with VectorEnv._add_info.

action_space

action_space

observation_space

observation_space

single_action_space

single_action_space

single_observation_space

single_observation_space