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

Signed-off-by: Mengqing Cao <cmq0113@163.com>
2025-05-15 14:16:31 +08:00
parent dd2a94596a
commit 70f8b96724
1 changed files with 1 additions and 0 deletions
--- a/vllm/model_executor/layers/fused_moe/layer.py
+++ b/vllm/model_executor/layers/fused_moe/layer.py
@ -43,6 +43,7 @@ if current_platform.is_cuda_alike():
        from .pplx_prepare_finalize import PplxPrepareAndFinalize
 else:
    fused_experts = None  # type: ignore
+    FusedMoEPrepareAndFinalize = None  # type: ignore
 if is_rocm_aiter_moe_enabled():
    from vllm.model_executor.layers.fused_moe.rocm_aiter_fused_moe import (  # noqa: E501
        rocm_aiter_biased_group_topk as grouped_topk)