Getting empty response from Mixtral-8x7b with Llamaindex

Question

I am getting a empty response when querying my vector database with using Mixtral-8x7b with llamaindex.

I would like to know if the parameters and arguments is the right way to initialize the Mixtral LLM model. Please advise

def completion_to_prompt(completion):
  return f"[INST] {completion} [/INST]"

def messages_to_prompt(messages):
  messages_str ="/n".join([str(x) for x in messages])
  return f"[INST] {messages_str} [/INST] "

llm = HuggingFaceLLM(
            model_name="TheBloke/Mixtral-8x7B-v0.1-GPTQ",
            tokenizer_name="TheBloke/Mixtral-8x7B-v0.1-GPTQ",
            query_wrapper_prompt=PromptTemplate("[INST] {query_str} [/INST]"),
            context_window=3900,
            max_new_tokens=512,
            # tokenizer_kwargs={},
            generate_kwargs={"temperature": 0.1, "top_k": 50, "top_p": 0.7, "do_sample":True},
            messages_to_prompt=messages_to_prompt,
            completion_to_prompt=completion_to_prompt,
            device_map="auto",
        )