Crazy Easy OpenAI Fine Tuning

No time for long posts? No worries I got you.

Let's get started!

1. Define your use case

I decided on a "chatbot for a hotel with crazy rules".

hotelrules.txt

2. Generate fine-tuning formatted data for your use case

"Create [NUMBER] user chat samples that [USE CASE DESCRIPTION], the sample must be in openai finetuning data jsonl format"

hotelruleschatdata.jsonl

3. Log into OpenAI's Platform

Select Dashboard -> Fine-tuning -> Create

If you don't have an account yet, check this link.

4. Congrats you just fine-tuned!

Now to test your model just click "Playground" on the bottom right.

Look at the comparison between the base model and the fine-tuned one:

The fine-tune model knows to tell the user to wear a costume when using the elevator

That’s all folks! My objective was to help you get started as fast as possible, there are tons of other things to try out, and understand!

Give it some claps to make others find it, too! Also, Make sure you follow me on Medium to not miss anything. Let’s connect on LinkedIn.