seed¶
Functions
The worker_init_fn that Lightning automatically adds to your dataloader if you previously set set the seed with |
|
Reset the seed to the value that |
|
Function that sets seed for pseudo-random number generators in: pytorch, numpy, python.random In addition, sets the following environment variables: |
Helper functions to help with reproducibility of models.
-
pytorch_lightning.utilities.seed.
pl_worker_init_function
(worker_id, rank=None)[source]¶ The worker_init_fn that Lightning automatically adds to your dataloader if you previously set set the seed with
seed_everything(seed, workers=True)
. See also the PyTorch documentation on randomness in DataLoaders.
-
pytorch_lightning.utilities.seed.
reset_seed
()[source]¶ Reset the seed to the value that
pytorch_lightning.utilities.seed.seed_everything()
previously set. Ifpytorch_lightning.utilities.seed.seed_everything()
is unused, this function will do nothing.- Return type
-
pytorch_lightning.utilities.seed.
seed_everything
(seed=None, workers=False)[source]¶ Function that sets seed for pseudo-random number generators in: pytorch, numpy, python.random In addition, sets the following environment variables:
PL_GLOBAL_SEED: will be passed to spawned subprocesses (e.g. ddp_spawn backend).
PL_SEED_WORKERS: (optional) is set to 1 if
`workers=True
.
- Parameters
seed¶ (
Optional
[int
]) – the integer value seed for global random state in Lightning. If None, will read seed from PL_GLOBAL_SEED env variable or select it randomly.workers¶ (
bool
) – if set toTrue
, will properly configure all dataloaders passed to the Trainer with aworker_init_fn
. If the user already provides such a function for their dataloaders, setting this argument will have no influence. See also:pl_worker_init_function()
.
- Return type