News for hackers. Stuff that matters.
Re:How Google Translate squeezes deep learning onto a phone by ibrahima on Wednesday July 29, 2015 @03:03PM (#9970837)
Completely right. Applying a neural network is much faster than training one. The main trick here is fitting the trained model into cache (or smaller) so that the matrix multiplies are fast.

There are no comments on this comment yet.