Configured llm with together api provider to handle prompts.

This seems impossible to configure. AIMessage exception being returned, replace chatgpt with Qwen models.

initial commit

Duplicate from gradio-templates/chatbot