By definition, gradient descent takes infinitesimal steps. In other words, it's a maximally inefficient way to learn.
-
Your daily trolls are usually of higher quality
I'm making a serious point. Gradient descent is intrinsically inefficient because it takes very small steps (and if it doesn't, it's not GD).
-
-
It's such a broad statement that's hard to agree/disagree
Innneficient for what, mimizing functions? What class of functions? Among what class of optimizers? Is it related to your deep-learning posts? What if I could directly find the global minimum of my deep learning loss?Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.