I want to upload my pdf book for speciality question
1 Like
Hey! Today, you need to take your data (in this case a PDF) and convert it into a format that is compatible with the way the models are trained (prompt / completion pairs). We talk more about this in the docs: Fine-tuning with the Gemini API | Google for Developers
Broadly though, if you just have a PDF of a book, you should likely use RAG and not fine-tuning.
3 Likes
Hello. I just started using the latest version of Gemini 1.5 Pro by clicking buttons and reading the pop-ups. The pop-ups tell you what you can send or share during the session. I didn’t have much to share, so I shared a bunch of little things. I think the Model followed me well enough, even though I was using a combo of stuff for what I think is a pretty complex task. That said, try using the Audio Recording. I’m not sure, but it seemed to like that method. I’m just learning about all this, so I can’t tell you how to count the tokens; recording your own voice might be expensive.