Can anyone validate that they had success fine tuning with functions?

Can one person say they had some success with fine tuning functions on a 3.5-turbo model?

I haven’t read any success stories about it, and despite spending days on it, I am not making progress.

If you were successful please share details like: schema used, samples used, hyperparameters, etc.

If you were unsuccessful please also share that.

Other posts:

1 Like

Hey there! Any update on this? Currently working on a fine-tuned model for gpt-3.5 and gpt-4 with function calling. Wonder if it worked out for you!