Add context

#1
by Javier-Jimenez99 - opened

Hi! 😊

I am trying to train an embedding model to improve RAG. I have seen your mtg-rules dataset, which I suppose you used to create mtg-rules-qa. It would be incredible to have the rule used to generate each question/answer pair.

I have seen your work on MTG, and it looks amazing! 🔥 I am trying something similar but using more data and a curated dataset such as rulesguru.net. If you are still interested in creating this kind of LLM, I would love to collaborate with you.

Have a nice day!

I am also working on this, although independently of this dataset. If there is interest, do you want to start working together on a platform like Discord?

Sure this is my user: tripl3j

I'm twilightwolf90. I just sent you a friend request.

hello, sorry i didn't see your messages. i'll add you both on discord

@Javier-Jimenez99 looks like something is wrong with the metadata, that's what is supposed to have been used to generate the record but it looks wrong. ill see if i can figure out something

Sign up or log in to comment