Cross Fire

本来想自己用react实现一个，但是看到已经有大佬在搞了，自己也没时间，所以把自己思考的产品原型分享一下。 1. 参考openai官网的playground，可以在会话级实现调整接口返回temperature ，这样可以控制gpt创造性。 2. 增加预置promot的能力，可以针对会话预置promot，针对翻译等场景一键开启专用会话场景

perhaps system prompt should be session related not global

4

great project. I was thinking build something like this by myself, just don't have time to do it. in my original plan, I think system prompt should be a session...

perhaps the server part should use python instead of nodejs

2

I have investigate some python lib like `langchain` and `gpt_index`, it could create something like new bing or even more powerful function. perhaps the server part should be implement by...

another dynamic limit need

9

I have multiple resources that are being requested, each with varying capacities. Therefore, I need to implement rate limiting based on the capacity of each individual resource.

思路很厉害，不过我很好奇为什么对话可以这么长？

11

openai的对话长度应该只有4096个token而已，很难完成特别大的工程，一般都是对小的函数进行一下修改。为什么你可以连着聊那么多，chatgpt还能记住上下文呢？

the current version does not implements image reading and stream tool use , so I add it