Skip to content

Fix a crash in NeMo 2.0 during module._apply(lambda t: t.cpu()) #1502

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
timmoon10 merged 6 commits into from
Feb 25, 2025
Merged

Fix a crash in NeMo 2.0 during module._apply(lambda t: t.cpu()) #1502

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
fd76d53
Fix a crash with module._apply(lambda t: t.cpu())
guyueh1 Feb 22, 2025
f6ccf80
Add comments
guyueh1 Feb 22, 2025
e70f913
Merge branch 'main' into fix_crash_with_nemo_lightning
guyueh1 Feb 22, 2025
2e35ba6
Make sure tensor is moved to dst device before quantizer quantizes
guyueh1 Feb 24, 2025
05415c1
Merge branch 'main' into fix_crash_with_nemo_lightning
guyueh1 Feb 24, 2025
5df1dfe
Merge branch 'main' into fix_crash_with_nemo_lightning
timmoon10 Feb 25, 2025
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
2 changes: 2 additions & 0 deletions transformer_engine/pytorch/tensor/float8_tensor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -484,6 +484,8 @@ def _set_data(self, tensor: torch.Tensor) -> None:

# Tensor device
new_device = tensor.device if tensor.is_cuda else self.device
if not devices_match(new_device, tensor.device):
tensor = tensor.to(device=new_device)

# Just copy FP8 data if other tensor is Float8Tensor
if isinstance(tensor, Float8Tensor):
Expand Down
2 changes: 2 additions & 0 deletions transformer_engine/pytorch/tensor/mxfp8_tensor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -368,6 +368,8 @@ def _set_data(self, tensor: torch.Tensor) -> None:

# Tensor device
new_device = tensor.device if tensor.is_cuda else self.device
if not devices_match(new_device, tensor.device):
tensor = tensor.to(device=new_device)

# Just copy FP8 data if other tensor is MXFP8Tensor
if isinstance(tensor, MXFP8Tensor):
Expand Down