Skip to content

[MoE Layer] Fix moe_use_fusion_node Setting when EP==1 #722

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
hushenwei2000 wants to merge 1 commit into
base: develop
Choose a base branch
from
+2 −0
Open

[MoE Layer] Fix moe_use_fusion_node Setting when EP==1 #722

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
2 changes: 2 additions & 0 deletions src/paddlefleet/transformer/moe/moe_layer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -198,6 +198,8 @@ def __init__(
"moe_grouped_gemm is only supported when moe_token_dispatcher_type is 'deepep' and on GPU architecture SM90 or higher. If these conditions are not met, please set it to false in the configuration yaml."
)
self.fp8_dispatch = False
else:
self.moe_use_fusion_node = config.moe_use_fusion_node

if self.fp8:
if paddle.version.cuda() == "12.6":
Expand Down
Loading