News for hackers. Stuff that matters.

A comment on Consistency LLM: converting LLMs to parallel decoders accelerates inference 3.5x. Parent comment

Re:Consistency LLM: converting LLMs to parallel decoders acc... by PeterisP on Thursday May 09, 2024 @03:23PM (#40312002)
The models are multiple orders of magnitude smaller than the compressed versions of their training data, they can not be the equivalent of a DB of it.
[ Parent | on HN ]
Re:Consistency LLM: converting LLMs to parallel decoders acc... by lainga on Thursday May 09, 2024 @06:35PM (#40313623)
The training data is ideo-semantically compressed? News to me... is it perhaps stored in kanji?
[ Parent | on HN ]