Hacker News new | ask | show | jobs
by ilaksh 1248 days ago
Apparently each query requires hundreds of GBs of GPU RAM on several expensive accelerator cards.

Is the H100 deployed at Azure? I wonder how much more efficient that would be over A100s.
1 comments
jejeyyy77 1248 days ago
Which seems insane considering stable diffusion can run on a M1 MacBook.
link
ilaksh 1248 days ago
Sure but they are totally different algorithms doing different things.
link
jejeyyy77 1237 days ago
We’ll yeah
link