seed¶
Functions
A context manager that resets the global random state on exit to what it was before entering. |
Utilities to help with reproducibility of models.
- lightning.pytorch.utilities.seed.isolate_rng(include_cuda=True)[source]¶
A context manager that resets the global random state on exit to what it was before entering.
It supports isolating the states for PyTorch, Numpy, and Python built-in random number generators.
- Parameters
include_cuda¶ (
bool
) – Whether to allow this function to also control the torch.cuda random number generator. Set this toFalse
when using the function in a forked process where CUDA re-initialization is prohibited.
Example
>>> import torch >>> torch.manual_seed(1) <torch._C.Generator object at ...> >>> with isolate_rng(): ... [torch.rand(1) for _ in range(3)] [tensor([0.7576]), tensor([0.2793]), tensor([0.4031])] >>> torch.rand(1) tensor([0.7576])