How Fine-Tuning an OpenAI GPT Model with Your Own Data?

After OpenAI released GPTs on DevDay, thousands of Custom GPTs were pumped overnight.

But honestly, most of them are bullshit. They are just some custom prompts.

If everyone can easily create their own custom GPTs, why do we need a GPTs store?

If you want to be an entrepreneur, remember this rule:

Higher barriers, deeper moats;

lower barriers, shallower moats.

So, how to make your GPTs outstanding?

The answer is: to fine-tune your own model.

Here I want to share my experience fine-tuning my GPT model based on OpenAI API.

1. Define your Goal

Here, I want to start with a pet project. In this project, I hope to train a model that can determine the publication to which an article belongs based on its title.

We know, there are many publications on the Medium platform. Each publication has its own purpose.

• Better Programming focuses on programming-related articles
• Towards Data Science focuses on data science
• The Startup focuses on entrepreneurship.

Which publication does this article belong to?