Skip to content

[PyTorch debug] Fix issue with tp_group=None #2733

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
pggPL wants to merge 1 commit into
base: main
Choose a base branch
from
+7 −1
Open

[PyTorch debug] Fix issue with tp_group=None #2733

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
8 changes: 7 additions & 1 deletion transformer_engine/debug/features/utils/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,13 @@ def get_reduction_params(tensor_name: str, tp_group: torch.distributed.ProcessGr
reduce_within_microbatch = tensor_name != "weight"
if tensor_name == "weight":
if TEDebugState.weight_tensor_tp_group_reduce:
reduction_group = tp_group
# Do not overwrite with `None`: in torch.distributed collectives
# group=None means the default/world process group.
if tp_group is not None:
reduction_group = tp_group
else:
# "Reduce in TP group" requested, but TP group is missing.
skip_reduction = True
else:
skip_reduction = True
return skip_reduction, reduction_group, reduce_within_microbatch
Expand Down
Loading