diff --git a/src/llamafactory/chat/vllm_engine.py b/src/llamafactory/chat/vllm_engine.py
index 21f09a58..15ffd047 100644
--- a/src/llamafactory/chat/vllm_engine.py
+++ b/src/llamafactory/chat/vllm_engine.py
@@ -173,7 +173,7 @@ class VllmEngine(BaseEngine):
             multi_modal_data = None
 
         result_generator = self.model.generate(
-            inputs={"prompt_token_ids": prompt_ids, "multi_modal_data": multi_modal_data},
+            prompt={"prompt_token_ids": prompt_ids, "multi_modal_data": multi_modal_data},
             sampling_params=sampling_params,
             request_id=request_id,
             lora_request=self.lora_request,