Skip to content

GitLab

Unverified Commit 70f8b967 authored by Mengqing Cao's avatar Mengqing Cao Committed by GitHub
Browse files

[Bugfix] Fix FusedMoEPrepareAndFinalize for cuda-disalike backends (#18178) 


Signed-off-by: default avatarMengqing Cao <cmq0113@163.com>
parent dd2a9459
Hide whitespace changes
Inline Side-by-side
Showing with 1 addition and 0 deletions
vllm/model_executor/layers/fused_moe/layer.py
View file @ 70f8b967
...@@ -43,6 +43,7 @@ if current_platform.is_cuda_alike(): ...@@ -43,6 +43,7 @@ if current_platform.is_cuda_alike():
from .pplx_prepare_finalize import PplxPrepareAndFinalize from .pplx_prepare_finalize import PplxPrepareAndFinalize
else: else:
fused_experts = None # type: ignore fused_experts = None # type: ignore
FusedMoEPrepareAndFinalize = None # type: ignore
if is_rocm_aiter_moe_enabled(): if is_rocm_aiter_moe_enabled():
from vllm.model_executor.layers.fused_moe.rocm_aiter_fused_moe import ( # noqa: E501 from vllm.model_executor.layers.fused_moe.rocm_aiter_fused_moe import ( # noqa: E501
rocm_aiter_biased_group_topk as grouped_topk) rocm_aiter_biased_group_topk as grouped_topk)
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment