Great!!

I'm working on MindPal - A platform to build AI Agents and Multi-agent workflow to automate tasks for business.

It will just need a minute to build a multi-agent workflow on MindPal. If you are having complicated AI processes that could not be done by ChatGPT, give it a go at https://mindpal.space/

I saw another repo use weights from llama3 to construct embedding model. I think the use case will be use this small model to search and use bigger model to re-rank later. So the question will be how this approach compared to BM25

Hey, you are thinking about wordllama. We explicitly compare to wordllama in the results. Feel free to check it out.

Your point about BM25 is very interesting. We can run some comparisons! Thanks

Sorry for being stupid, but I have a python script to index website too. But I dont understand why it's alway can not index pages on time. How your tool can solve it

Since data used to train chatgpt is public internet data, it will probable be bad at anything uncommon niche knowledge