reference/generated/torchrl.trainers.algorithms.configs.utils.RpropConfig

Run in Google Colab

Colab

Download Notebook

Notebook

View on GitHub

GitHub

torchrl.trainers.algorithms.configs.utils.RpropConfig

class torchrl.trainers.algorithms.configs.utils.RpropConfig(lr: float = 0.01, etas: tuple[float, float] = (0.5, 1.2), step_sizes: tuple[float, float] = (1e-06, 50.0), capturable: bool = False, foreach: bool | None = None, maximize: bool = False, differentiable: bool = False, _target_: str = 'torch.optim.Rprop', _partial_: bool = True)

Hydra configuration for torch.optim.Rprop.

Every kwarg accepted by torch.optim.Rprop.__init__ is exposed as a field here.