Hi, how can I ask questions based on historical conversations?

[bswaterb]

I found the dialogue history JSON in the data/dialogue directory. Now, I want to load this data as my conversation history and ask a new question to test the performance of the SCM method. How can I do this? Thanks for your help!

[wbbeyourself]

Just put the dialogue id in the input box, and press enter, the model will automatic load the data and construct embedding, but it will take some time.

[wbbeyourself]

请问同学你是从哪看到我这篇论文的呀？

[wbbeyourself]

Just put the dialogue id in the input box, and press enter, the model will automatic load the data and construct embedding, but it will take some time.

Run dialogue_test.py, and put the dialogue id in the first input.

[bswaterb]

哈哈哈，我有做类似多轮对话上下文管理框架的idea，来研究一下大佬是怎么做的工作，感谢答疑～

[wbbeyourself]

哈哈哈，我有做类似多轮对话上下文管理框架的idea，来研究一下大佬是怎么做的工作，感谢答疑～

哈哈哈哈 感谢使用哈，代码当时写的有点简陋，就是想用一个方便的demo测试方法的效果的，最近忙新的研究了，也很久没更新了。有啥问题随时提哈。现在调 api和embedding 的方式不知道是否失效，openai确实更新了很多。

[bswaterb]

哈哈哈，我有做类似多轮对话上下文管理框架的idea，来研究一下大佬是怎么做的工作，感谢答疑～

哈哈哈哈 感谢使用哈，代码当时写的有点简陋，就是想用一个方便的demo测试方法的效果的，最近忙新的研究了，也很久没更新了。有啥问题随时提哈。现在调 api和embedding 的方式不知道是否失效，openai确实更新了很多。

好滴，有问题我再咨询，再次感谢大佬