Fine-tuning on the insertion task?

Hi all, I’d like to fine-tune davinci on text insertions rather than completions, i.e. my training data would ideally include prompt, suffix, and completion. Is this currently supported?

I ask because it appears that the Python openai.Completion.create method does not currently support the suffix argument when the model provided is a fine-tuned davinci model. (I get the error: Unrecognized request argument supplied: suffix). But perhaps I’m doing something wrong, or perhaps what I’m envisioning will work for other LLMs besides davinci. Thanks!

I’m having the same issue… Are there any updates on this?