I am building caption generator app using Gemini ai but I don’t understand how to fine tune model to response only in json. ?
Welcome to the forums!
You don’t show us the code you’ve tried, so it is difficult to give details, but in general you want to use structured output to format the results.
Currently, Gemini API supports finetuning for only text input and output. And also once you finetune the model, JSON mode is not supported.
1 Like