3 janv. 2015 - “Binaryconnect: Training deep neural networks with binary weights during propagations, NIPS
Description:
BC: sample the original weights into binary orternary according to the value of the weights (after scaling to [0,1]) as the sampling probability.
Ajouté au bande de temps:
Date:
3 janv. 2015
Maintenaint
~ Il y a 9 ans et 5 mois
