I want to create a separate fine tuned model for separate use cases like
ft-ada-user1, ft-ada-user-2 via using API. And then in the play ground under Model dropdown i should be able to select the Fine tuned model and chat with my data
is there any way to do it?
Multiple fine-tune models can be made.
Each model you create by fine-tuning has its own auto-generated name, and you can also add a prefix to make them easier to identify.
However “chat with your data” is not a good use case for fine-tune behavior training. You likely want to first investigate a vector database information retrieval system to give the AI relevant information to the user question.
ada is going away - and can barely “chat”. The new cheap completion model to fine-tune would be babbage-002.
1 Like
Thanks!
So removing Fine tuning out. Is it possible to create separate blank base models for different users. Each model should give me a separate End point to access.
Like 3 versions of same base model. gpt-3.5-turbo1, gpt-3.5-turbo2, gpt-3.5-turbo3.
ps:- Via API only
Before we get into a XY Problem, what’s your actual end goal for this project?
1 Like
If the question is if the fine tuned models will be available in the playground then the answer is “yes”.
PS. Those are not my models. I just looked this up in another post in the forum.
1 Like
1 Like