[Issue]: Implementing Conversational Capabilities in UserProxyAgent

[wbw625]

Describe the issue

I am currently working with ConversableAgent and UserProxyAgent classes. I am using a local model Mistral-7B-OpenOrca for my implementation. I noticed that the ConversableAgent can perform Q&A functions, but the UserProxyAgent primarily executes code.

I want to extend the functionality of UserProxyAgent so that it can also perform Q&A similar to ConversableAgent. Here is the current setup I am using:

assistant = ConversableAgent("agent", llm_config={"config_list": config_list})

user_proxy = UserProxyAgent(
    "user_proxy",
    code_execution_config={"work_dir": "coding", "use_docker": False},
    # llm_config={"config_list": config_list}
    llm_config=False
)

assistant.register_model_client(model_client_cls=CustomModelClient)
# user_proxy.register_model_client(model_client_cls=CustomModelClient)
user_proxy.initiate_chat(assistant, message="Plot a chart of NVDA and TESLA stock price change YTD.")

My question is: How can I make UserProxyAgent capable of performing Q&A functionalities, and not just running code?

Thank you.

Steps to reproduce

No response

Screenshots and logs

No response

Additional Information

The configuration file OAI_CONFIG_LIST is as follows:

[
    {
        "model": "Open-Orca/Mistral-7B-OpenOrca",
        "model_client_cls": "CustomModelClient",
        "device": "cuda:3",
        "n": 1,
        "params": {
            "max_length": 1000
        }
    }
]