Cannot use functions when deploying local model

[choiszt]

I am currently deploying the LLaVA series based on the qwen-agent framework. However, I encountered an issue while trying the basic demo, where I am unable to get any return for function calls. It appears that the code does not take functions as input, and as a result, the _call_llm method does not return any function entry points or outputs.

After investigating the issue, I found that the _chat_with_functions method doesn’t seem to handle function calls properly. Below is the relevant section of the code:
def _chat_with_functions( self, messages: List[Message], functions: List[Dict], stream: bool, delta_stream: bool, generate_cfg: dict, lang: Literal['en', 'zh'], ) -> Union[List[Message], Iterator[List[Message]]]: if delta_stream: raise NotImplementedError('Please use stream=True with delta_stream=False, because delta_stream=True' ' is not implemented for function calling due to some technical reasons.') generate_cfg = copy.deepcopy(generate_cfg) for k in ['parallel_function_calls', 'function_choice']: if k in generate_cfg: del generate_cfg[k] return self._continue_assistant_response(messages, generate_cfg=generate_cfg, stream=stream)
Is there any solutions?