# If teach a finetune-id a set of information and that information changes over time let's say 'twitter posts', how can I go back and train my fine-tune-id on the updated data