I am trying to train a GPT model to answer A level questions like the students in my class, but it doesnt have the knowledge required. I know I can train two seperate bots for this but ideally I would like to have it done all at once.
Is it possible?
Welcome to the community!
You can either:
A. fine-tune GPT 3.5 through the API, and create your own assistant that uses the fine-tuned model’s API calls, or
B. provide documents inside a custom GPT as “knowledge”, so the custom GPT can retrieve the relevant information.
Whichever choice you make depends on your needs and level of expertise.
Hello! Thanks for answering, but I was wondering if it was possible to do both embed and fine tune in one bot like Llama v2.
Like, allow a fine-tuned model access to specific vector embeddings? Then yes, if you use the APIs. I would research vector databases in that case.