torchrl.trainers.algorithms.configs.data.LazyMemmapStorageConfig¶
- class torchrl.trainers.algorithms.configs.data.LazyMemmapStorageConfig(_partial_: bool = False, _target_: str = 'torchrl.data.replay_buffers.LazyMemmapStorage', max_size: int | None = None, scratch_dir: Any = None, device: Any = 'cpu', ndim: int = 1, existsok: bool = False, compilable: bool = False, shared_init: bool = False, auto_cleanup: bool | None = None)[source]¶
Hydra configuration for
LazyMemmapStorage.Every kwarg accepted by
LazyMemmapStorage.__init__is exposed as a field here.