reference/generated/torchrl.trainers.algorithms.configs.objectives.GAEConfig

Run in Google Colab

Colab

Download Notebook

Notebook

View on GitHub

GitHub

torchrl.trainers.algorithms.configs.objectives.GAEConfig

class torchrl.trainers.algorithms.configs.objectives.GAEConfig(_partial_: bool = False, gamma: float | None = None, lmbda: float | None = None, value_network: Any = None, average_gae: bool = True, differentiable: bool = False, vectorized: bool | None = None, skip_existing: bool | None = None, advantage_key: str | None = None, value_target_key: str | None = None, value_key: str | None = None, shifted: bool = False, device: Any = None, time_dim: int | None = None, auto_reset_env: bool = False, deactivate_vmap: bool = False, value_chunk_size: int | None = None, num_chunks: int | None = None, num_chunk: int | None = None, shifted_budget: int = 1, _target_: str = 'torchrl.objectives.value.GAE')

Hydra configuration for GAE.

Every kwarg accepted by GAE.__init__ is exposed as a field here.