Hacker News new | ask | show | jobs
by hnfong 798 days ago
With quantized models you can run 70B models on 64GB RAM comfortably.