#Training Information
6 messages · Page 1 of 1 (latest)
You can use embedding to train and search your query. I am working on a similar project DM if you have questions.
I'd like to know more about this method as well. I am able to generate embedding vectors but then what do I do with them besides throw them in a db or cosine similarity...
I recently tried making a fine tuning model but it's not inspecting it so much as it's modeling it to simulate (inaccurate) responses. I want to be able to put a large corpus of text into the AI (or my app) and ask it questions about it that may require broad fine-grained knowledge.
Yes, I have a similar issue. I would like to set up a data model, and then later generate SQL queries without having to specify the data model again and again. I am not clear on how embeddings are supposed to help me in that.