How do you call the APIs? With GPT's function calling or your own design?
Hi, I was reading your paper(great paper by the way). And I didn't see the details about how you were calling the APIs. Could you be so kind to tell me?
@caizhuoyue77 Thanks! Model calling is embedded in the LLMNode and you can easily replace it with your own models. eg. if you are using OAI API, https://github.com/billxbf/ReWOO/blob/9cd0283043ff4be0c9d614fda2789d143ca6ffd1/nodes/LLMNode.py#L48
You may need to update it a bit since it's a very early work. Let me know if you need help.
Thanks for replying. But I tried your script and it can ever answer the most simple question. I am confused. I configured the keys and use GPT-4 as model and ask some question about the new Prime Minister of Japan. It refused to answer and says it cannot predict the future.