[infer] fix vllm args (#7235)

Former-commit-id: 999be5b4512890b8cf4f45874a77e35cf35626f5

src/llamafactory/chat/vllm_engine.py

@@ -170,7 +170,7 @@ class VllmEngine(BaseEngine):
             or 1.0,  # repetition_penalty must > 0
             temperature=temperature if temperature is not None else self.generating_args["temperature"],
             top_p=(top_p if top_p is not None else self.generating_args["top_p"]) or 1.0,  # top_p must > 0
-            top_k=top_k if top_k is not None else self.generating_args["top_k"],
+            top_k=(top_k if top_k is not None else self.generating_args["top_k"]) or -1,  # top_k must > 0
             stop=stop,
             stop_token_ids=self.template.get_stop_token_ids(self.tokenizer),
             max_tokens=max_tokens,