Class VectorRewardWrapper<TObs, TAct>
A vector wrapper that changes only the rewards, through Rewards(NDArray), after every step.
public abstract class VectorRewardWrapper<TObs, TAct> : VectorWrapper<TObs, TAct, TObs, TAct>, IVectorEnv, IDisposable
Type Parameters
TObsThe batched observation type.
TActThe batched action type.
- Inheritance
-
VectorEnv<TObs, TAct>VectorWrapper<TObs, TAct, TObs, TAct>VectorRewardWrapper<TObs, TAct>
- Implements
- Derived
- Inherited Members
- Extension Methods
Remarks
Reference: gymnasium/vector/vector_env.py.
Constructors
VectorRewardWrapper(VectorEnv<TObs, TAct>)
Wraps env.
protected VectorRewardWrapper(VectorEnv<TObs, TAct> env)
Parameters
envVectorEnv<TObs, TAct>The vector env to wrap.
Exceptions
Methods
Rewards(NDArray)
The transform (rewards(rewards)).
public abstract NDArray Rewards(NDArray rewards)
Parameters
rewardsNDArrayThe env's
(num_envs,)rewards.
Returns
- NDArray
The wrapper's rewards.
Step(TAct)
Steps the env and transforms its rewards.
public override VectorStepResult<TObs> Step(TAct actions)
Parameters
actionsTActThe actions.
Returns
- VectorStepResult<TObs>
The transition with transformed rewards.