Skip to content

WIP: Disable _optimizer_to_device logic #20036

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: master
Choose a base branch
from
Open

WIP: Disable _optimizer_to_device logic #20036

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
137c7b9
Disable _optimizer_to_device logic
corwinjoy Jul 2, 2024
d2589fc
update test
awaelchli Jul 3, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 1 addition & 5 deletions src/lightning/fabric/utilities/optimizer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,11 +14,8 @@

from typing import Iterable

from lightning_utilities.core.apply_func import apply_to_collection
from torch import Tensor
from torch.optim import Optimizer

from lightning.fabric.utilities.apply_func import move_data_to_device
from lightning.fabric.utilities.types import _DEVICE


Expand All @@ -30,5 +27,4 @@ def _optimizers_to_device(optimizers: Iterable[Optimizer], device: _DEVICE) -> N

def _optimizer_to_device(optimizer: Optimizer, device: _DEVICE) -> None:
"""Moves the state of a single optimizer to the device."""
for p, v in optimizer.state.items():
optimizer.state[p] = apply_to_collection(v, Tensor, move_data_to_device, device, allow_frozen=True)
pass
2 changes: 1 addition & 1 deletion tests/tests_pytorch/trainer/test_trainer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1766,7 +1766,7 @@ def current_memory():
trainer.fit(model)

assert trainer.strategy.model is model
assert list(trainer.optimizers[0].state.values())[0]["exp_avg_sq"].device == torch.device("cpu")
assert list(trainer.optimizers[0].state.values())[0]["exp_avg_sq"].device == torch.device("cuda", 0)
assert trainer.callback_metrics["train_loss"].device == torch.device("cpu")

assert current_memory() <= initial
Expand Down