RLEnv class
Reinforcement Learning (RL) environment class which subclasses gym.Env.
This is a wrapper over Env for RL users. To create custom RL environments users should subclass RLEnv and define the following methods: get_reward_range(), get_reward(), get_done(), get_info().
As this is a subclass of gym.Env, it implements reset() and
step().
Methods
- def close(self) -> None
- def current_episode(self, all_info: bool = False) -> dataset.BaseEpisode
- Returns the current episode of the environment.
- def get_done(self, observations: simulator.Observations) -> bool
- Returns boolean indicating whether episode is done after performing the last action.
- def get_info(self, observations) -> typing.Dict[typing.Any, typing.Any]
- def get_reward(self, observations: simulator.Observations) -> typing.Any
- Returns reward after action has been performed.
- def get_reward_range(self)
- Get min, max range of reward.
- def render(self, mode: str = 'rgb') -> numpy.ndarray
- def reset(self, *, return_info: bool = False, **kwargs) -> typing.Union[simulator.Observations, typing.Tuple[simulator.Observations, typing.Dict]]
- def seed(self, seed: typing.Optional[int] = None) -> None
- def step(self, *args, **kwargs) -> typing.Tuple[simulator.Observations, typing.Any, bool, dict]
- Perform an action in the environment.
Special methods
- def __class_getitem__(...)
- Parameterizes a generic class.
- def __enter__(self)
- def __exit__(self, exc_type, exc_val, exc_tb)
- def __format__(self, format_spec, /)
- Default object formatter.
- def __init__(self, config: DictConfig, dataset: typing.Optional[dataset.Dataset] = None)
- Constructor
- def __init_subclass__(...)
- Function to initialize subclasses.
- def __str__(self)
Properties
- config: DictConfig get
- episodes: typing.List[dataset.Episode] get set
- habitat_env: Env get
- np_random: gym.utils.seeding.RandomNumberGenerator get set
- Initializes the np_random field if not done already.
- unwrapped: gym.core.Env get
- Completely unwrap this env.
Data
- metadata = {'render_modes': []}
- reward_range = (-inf, inf)
- spec = None
Method documentation
def habitat.
Returns the current episode of the environment.
| Parameters | |
|---|---|
| all_info | If true, all the information in the episode will be provided. Otherwise, only episode_id and scene_id will be included. |
| Returns | The BaseEpisode object for the current episode. |
def habitat.
Returns boolean indicating whether episode is done after performing the last action.
| Parameters | |
|---|---|
| observations | observations from simulator and task. |
| Returns | done boolean after performing the last action. |
This method is called inside the step method.
def habitat.
| Parameters | |
|---|---|
| observations | observations from simulator and task. |
| Returns | info after performing the last action. |
def habitat.
Returns reward after action has been performed.
| Parameters | |
|---|---|
| observations | observations from simulator and task. |
| Returns | reward after performing the last action. |
This method is called inside the step() method.
def habitat.
Get min, max range of reward.
| Returns | [min, max] range of reward. |
|---|
def habitat.
Perform an action in the environment.
| Returns | (observations, reward, done, info) |
|---|
def habitat.
Parameterizes a generic class.
At least, parameterizing a generic class is the main thing this method does. For example, for some generic class Foo, this is called when we do Foo[int] - there, with cls=Foo and params=int.
However, note that this method is also called when defining generic classes in the first place with class Foo[T]: ….
def habitat.
Default object formatter.
Return str(self) if format_spec is empty. Raise TypeError otherwise.
def habitat.
Constructor
| Parameters | |
|---|---|
| config | config to construct Env |
| dataset | dataset to construct Env. |
Property documentation
habitat.
Completely unwrap this env.
- Returns:
- gym.Env: The base non-wrapped gym.Env instance