News for hackers. Stuff that matters.
Re:Google employees criticize CEO for “dumpster fire” respon... by visarga on Tuesday February 14, 2023 @03:04PM (#34795154)
Apparently Google is capable of doing cool things. DeepMind’s speculative sampling achieves 2–2.5x decoding speedups in LLM. That brings cost down significantly, without degradation in quality.

There are no comments on this comment yet.