I train a davinci model in Google Colab with the aim to reach a certain level of context dependence (training data are anonymized chat logs).
Is there a way to alter the output length? Out of the box it seems to me as there is no option to change the output length.
I use following structure:
First, I create the training data.
Then, I start the fine tuning:
!openai api fine_tunes.follow -i ft-xxxxxx
Finally, I can test the model with a prompt:
!openai api completions.create -m davinci:ft-personal-2022-09-12-08-32-19 -p “How are you doing?”
I appreciate any hints or thoughts on this.