The race in AI should be toward models with fewer and fewer parameters for a given performance, not the opposite.
-
They are impractically large when you need to deploy them for 3 billion users.
-
-
They probably use a distilled version of BERT. But my point is that those large models a maybe impractical for real-time usage, but they are very practical to use as teachers for lighter models.
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
FB & Instagram use large multilingual transformers and large ConvNets for all kinds of things (content moderation, translation, ranking, etc). But there are large teams of engineers who make them as efficient as possible.
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.