| License | MIT |
|---|---|
| Stability | experimental |
| Portability | non-portable |
| Safe Haskell | None |
| Language | Haskell2010 |
OpenAI.Gym.Data
Description
Aeson-based data types to be returned by OpenAI.Gym.API
- data GymEnv
- newtype InstID = InstID {}
- newtype Environment = Environment {}
- newtype Observation = Observation {}
- data Step = Step {}
- data Outcome = Outcome {}
- newtype Info = Info {}
- newtype Action = Action {}
- data Monitor = Monitor {}
- data Config = Config {
- training_dir :: !Text
- algorithm_id :: !Text
- api_key :: !Text
Documentation
Classic Control Environments
Constructors
| CartPoleV0 | Balance a pole on a cart (for a short time). |
| CartPoleV1 | Balance a pole on a cart. |
| AcrobotV1 | Swing up a two-link robot. |
| MountainCarV0 | Drive up a big hill. |
| MountainCarContinuousV0 | Drive up a big hill with continuous control. |
| PendulumV0 | Swing up a pendulum. |
| FrozenLakeV0 | Swing up a pendulum. |
| PongRamV0 | Maximize score in the game Pong, with RAM as input |
| PongV0 | Maximize score in the game Pong |
a short identifier (such as '3c657dbc') for the created environment instance. The instance_id is used in future API calls to identify the environment to be manipulated.
newtype Environment Source #
a mapping of instance_id to env_id (e.g. {'3c657dbc': 'CartPole-v0'}) for every env on the server
Constructors
| Environment | |
Instances
newtype Observation Source #
The agent's observation of the current environment
Constructors
| Observation | |
Fields | |
Instances
An action to take in the environment and whether or not to render that change
The result of taking a step in an environment
Constructors
| Outcome | |
A dict containing auxiliary diagnostic information
An action to take in the environment
Parameters used to start a monitoring session.
Constructors
| Monitor | |
Fields
| |
Parameters used to upload a monitored session to OpenAI's servers
Constructors
| Config | |
Fields
| |