Researchers upend AI status quo by eliminating matrix multiplication in LLMs

Re:Researchers upend AI status quo by eliminating matrix mul... by JKCalhoun on Tuesday June 25, 2024 @07:13PM (#40794775)

"Call my broker, tell him to sell all my NVDA!"

Combined with the earlier paper this year that claimed LLMs work fine (and faster) with trinary numbers (rather than floats? or long ints?) — the idea of running a quick LLM local is looking better and better.

[ Parent | on HN ]

Re:Researchers upend AI status quo by eliminating matrix mul... by cgearhart on Tuesday June 25, 2024 @07:22PM (#40794820)

This is the same paper (or an extension) — using ternary weights means you can replace multiplication with addition/subtraction.

[ Parent | on HN ]