all a / b / c / f / h / j / jp / l / o / q / s / sw / lounge cgi up wiki
[Settings] [Home] [Contact] [Catalog] [Search] [Thread list] [Stats] [Reports] [Watcher] [PMs] [Admin]
[Return]

Report a post

Preview
No.158791 [Report] +
▶
>>158788
>There's a reason they run whole datacentres just to be able to build the next flagship LLM, with terabytes of VRAM
I'm curious: what's the hardware and power needed to train (not run, train from scratch) a model for Stable Diffusion? Suppose I have 100 million images (somehow) and I want to build a local model. What do I need?

I do wonder if the reason they use entire data centers to build the next cutting edge LLM is because they're rushing to meet deadlines and can afford to be wasteful with computing hardware.

Your fortune: Average luck
Post number No.158791
Board Off-Topic@Heyuri
Optional. Describe what's wrong with it.