2026-11-13 · America/Los_Angeles · 社区动态 · #14
I made a free tool that tells you which local models your PC can actually run (and why)
When I started with local LLMs I spent a lot of time downloading models that didn't fit, or that ran at 2 tok/s because half of them ended up in RAM D: So I built Jaull for my final-year project. You run it, it checks your CPU, RAM and GPU, and asks a few plain questions: what you want to use it for, which language, whether you care more about speed or quality. Then it searches Hugging Face (metadata only, it doesn't download anything) and shows up to five models that fit your machine, with the exact file to get and why it picked each one. It also tells you which numbers are estimates and…
热度 45.4 / 100;排名与评分保留该期记录。