OakEmu-RL v0.1.0

Appearance

RewardShapingBatch ​

python
RewardShapingBatch(observations: torch.Tensor, actions: torch.Tensor | None = None, rewards: torch.Tensor | None = None, next_observations: torch.Tensor | None = None, dones: torch.Tensor | None = None, metadata: dict[str, torch.Tensor] | None = None, step: int = 0) -> None

A learner-side batch made available to reward shapers for batched updates.

Source

Values ​

actions ​

python
actions: torch.Tensor | None = None

Source

dones ​

python
dones: torch.Tensor | None = None

Source

metadata ​

python
metadata: dict[str, torch.Tensor] | None = None

Source

next_observations ​

python
next_observations: torch.Tensor | None = None

Source

observations ​

python
observations: torch.Tensor

Source

rewards ​

python
rewards: torch.Tensor | None = None

Source

step ​

python
step: int = 0

Source