Skip to content

Fix bitsandbytes layer conversion under init_module context manager #18914

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

Merged
merged 6 commits into from
Nov 1, 2023
Merged

Fix bitsandbytes layer conversion under init_module context manager #18914

Changes from 1 commit
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
0f58bd2
fix
awaelchli Nov 1, 2023
08e5180
docker
awaelchli Nov 1, 2023
6cd2f14
fabric
awaelchli Nov 1, 2023
cb4a0d0
try a different way
awaelchli Nov 1, 2023
524226b
Merge branch 'master' into bugfix/bnb-init
awaelchli Nov 1, 2023
147b4b3
chlog
awaelchli Nov 1, 2023
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: 4 additions & 2 deletions src/lightning/fabric/strategies/strategy.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -138,11 +138,13 @@ def module_init_context(self, empty_init: Optional[bool] = None) -> ContextManag
If ``None``, the strategy will decide. Some strategies may not support all options.

"""
tensor_init_ctx = self.tensor_init_context()
precision_module_ctx = self.precision.module_init_context()
stack = ExitStack()
if _TORCH_GREATER_EQUAL_2_0:
stack.enter_context(self.root_device)
if _TORCH_GREATER_EQUAL_1_13:
stack.enter_context(_EmptyInit(enabled=bool(empty_init)))
stack.enter_context(tensor_init_ctx)
stack.enter_context(precision_module_ctx)
return stack

def setup_module_and_optimizers(
Expand Down