OpenAI - Intro & Embeddings

[@SridharKumarKannam]

If you found this content useful, pleases consider sharing it with others who might benefit. Your support is greatly appreciated :)

[@nourinahmedeka9518]

This tutorial series should blow up! Thank you for putting it all together. I especially appreciate the brief overview of building concepts - bag of words, TF-IDF, and embeddings. Thanks!

[@SridharKumarKannam]

Thank you very much :)

[@hiteshgupta968]

That's a great tutorial I want to understand.... Does that really matter which model we use for embedding .... Logically it feels like we are generating floting number which is just a basic mathematical computation....what are your thoughts?

[@SridharKumarKannam]

yes, it does matter. For example, models which trained on specific dataset (for example medical) can't create good embeddings for legal data. Even though most models are trained on generic data, which can be used for any data, you can try a couple of models to check which one works best for you.

[@hiteshgupta968]

@@SridharKumarKannam so I am just thinking off can we use a multi model embedding approach to make production ready RAG .. for example I will use an model which is more specific to audio and another model for text data and store all of the embedding in vector database... What's your thoughts?