首次微調，需要理智檢查——3B 還是 7B 用於多任務推理？[D]

Ok so this is my first post here, been lurking for a while. I’m about to start my first fine-tuning project and I don’t want to commit to the wrong direction so figured I’d ask. Background on me: I’m not from an ML background, self-taught, been working with LLMs through APIs for about a year. Hit the wall where prompt engineering isn’t enough anymore for what I’m trying to do, so now I need to actually fine-tune something. Here’s the task. I want the model to learn three related things: First