Are you struggling with fine-tuning (possible solution)

General API discussion
1

I am looking to help one or two people with fine-tuning (at no charge)

If you struggle to get it to work effectively, can you send me a private chat message explaining your use case?

I may have a really simple solution - but I would like to check it works as a generic solution by helping a couple of people with real-world problems in a couple of different use cases. (I have already tested it with a small and large set of data - so I am fairly confident it works)

If it works well, I will tell everyone about it.

5 Likes
2

If people can’t figure out fine tuning from the docs they’re not cut out for this stuff.

AIML development isn’t for everyone

3

This is a rather arrogant comment.

There are 2 problems that users may encounter:

  1. Technical problem: All users don’t know python and have to either learn python or use rest api. The rest api is not as well documented in the documentation as python.

  2. Problem with data: Fine tuning is a very complex problem and if it wasn’t like that, all of us would have already created our own ChatGpt overnight.

I started using gpt3 a year ago and I would be grateful if someone offered me help like this. Of course today I have no more technical problems and my models are at least 10x better than at the beginning.

I’m lucky that you didn’t write this post a year ago :slight_smile:

4 Likes
4

GPT and AIML are two separate technologies. AIML is a markup code gpt uses python.

5

Anyone who thinks AIML is referencing a markup language is also not cut out for this stuff.

@sherylclyde go write some markup and I’ll go develop some deep reinforcement learning PyTorch systems.