Re:PowerInfer: Fast Large Language Model Serving with a Cons...
by voxl on Tuesday December 19, 2023 @06:37PM
(#38703511)
Probably, the idea that the GPU handles the "hot neurons" means that if your GPU is strong enough to handle _enough_ hot neurons, you should expect speedups.
[ Parent
| on HN ]
There are no comments on this comment yet.