diff --git a/vllm/model_executor/layers/fused_moe/experts/triton_moe.py b/vllm/model_executor/layers/fused_moe/experts/triton_moe.py index b9b046aa17..f3a8e40e54 100644 --- a/vllm/model_executor/layers/fused_moe/experts/triton_moe.py +++ b/vllm/model_executor/layers/fused_moe/experts/triton_moe.py @@ -237,7 +237,7 @@ class TritonExperts(LoRAExpertsMixin, mk.FusedMoEExpertsModular): hidden_states, w1, intermediate_cache1, - a1q_scale, + a1q_scale if a1q_scale is not None else self.a1_scale, self.w1_scale, None, # topk_weights sorted_token_ids,