News for hackers. Stuff that matters.
Re:Google Releases Paper on Embodied LLM (562b parameters) by ivalm on Tuesday March 07, 2023 @09:53PM (#35064608)
Chinchilla paper proposes the following empirical scaling power law

Loss= (N0/N)^a+(D0/D)^b

Where N is number of parameters, D is number of examples, a,b are exponents. N0,D0 are scaling factors.

There are no comments on this comment yet.