How developers are training OpenAI Codex models?

Hi there!

I read in the FAQ that at the moment, fine tuning the openAI Codex Models is not possible. But I see different tools now which for example are generating WordPress code with openAI. Here is only one of them: WPCodey - Your AI-powered WordPress Code Snippets Assistant

From the developer I read the following post some weeks/months ago: "I read about advanced AI and how to train a custom model. Today I got the chance to play a bit with it. OpenAI doesn’t know about Oxygen Builder, let alone about writing a custom Oxygen Element. So I tried to train it using some Oxygen Components and see if it can figure it out, it sure did!

How did he train the model? What possibilities we have here? Just for example to train a model to concentrate on WordPress PHP output for the code completion. I have searched the whole internet and find only info that this is currently not possible. :thinking:

What do you say? Thanks for your feedback!

1 Like

I don’t think there is a way to do this yet. I have yet to find anything current. There were older API’s available that I have seen referenced but those have been disabled.