Hacker News new | ask | show | jobs
by hiroakiaizawa 40 days ago
Nice. What scale does this realistically reach on a single machine?
1 comments
lynx97 40 days ago
Model: 36L/36H/576D, 144.2M params

runs on a Blackwell 6000 Max-Q, using 86GB VRAM. Training supposedly takes 3h40m

link