Follow

> In contrast to AdaGrad, they estimate the product of D and G in the denominator, so we call the proposed technique Prodigy

Product of D and GでProdigy。
なかなかいいセンス……いやそれでいいのか……?
最初見たときはちょっと脱力した。

arxiv.org/abs/2306.06101

Sign in to participate in the conversation
Mastodon

Experimental private instance. Running on FreeBSD!