diff --git a/vllm/model_executor/layers/fused_moe/fused_marlin_moe.py b/vllm/model_executor/layers/fused_moe/fused_marlin_moe.py index 8d95665f7..5d382cfc9 100644 --- a/vllm/model_executor/layers/fused_moe/fused_marlin_moe.py +++ b/vllm/model_executor/layers/fused_moe/fused_marlin_moe.py @@ -561,7 +561,7 @@ class MarlinExpertsBase(mk.FusedMoEPermuteExpertsUnpermute): @staticmethod def _supports_no_act_and_mul() -> bool: - return False + return True @staticmethod def _supports_quant_scheme(