Maintaining a fine-tuned model for every one of your users - even with technique... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		simonw on July 24, 2024 \| parent \| context \| favorite \| on: Solving the out-of-context chunk problem for RAG Maintaining a fine-tuned model for every one of your users - even with techniques like LoRA - sounds complicated and expensive to me!

lowdest on July 24, 2024 [–]

It is, but it's also not that bad. A copy of the weights is X GB of cloud storage, which can be stored as a diff if it helps, and added compute time for loading a custom model and unloading for the next customer. It's not free, but it's an approachable cost for a premium service.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact