Hacker News new | ask | show | jobs
by robonot 77 days ago
really impressive for the size. Curious to see what happens when someone trains a 100B+ model natively at 1-bit.