Would the following requirement be suitable for fine-tuning OpenAI's model?

API
1

The entire requirement includes:

  1. Contain the part name, brand name, and part category. (The part name and brand name determine the part category.)
  2. There are 520 fixed part categories, and I can provide 10 high-quality data for each category. In the future, after the business inputs the part name, the model will recommend a category.
2

In the fine-tuning docs on classification, you can see the full requirements, including that you should “aim for at least ~100 examples per class.”

You could fine-tune a model with 10 datapoints for each category, but it may not be very accurate.

3

Thank you for your response.
As you said, having more samples is essential for achieving more accurate results.
As a supply chain business professional, if I have enough sample data, it means that I have completed a specific business scenario. Currently, it is not very clear which specific business scenarios OpenAI can implement, especially for large enterprises.