Every flop counts: scaling a 300b mixture-of-experts Ling LLM without premium GPUs (arxiv.org)
from yogthos@lemmy.ml to technology@lemmy.ml on 21 May 21:42
https://lemmy.ml/post/30467866

#technology

threaded - newest

davel@lemmy.ml on 21 May 22:21 collapse

a 300 lb slurry of stack overflowers

ferric_carcinization@lemmy.ml on 22 May 06:50 collapse

Isn’t 300lb approximately equal to 1 stackoverflow user?

(I’m not USAian, so my conversion may have higher error bars than usual)