Compact ConvNets with Ternary Weights and Binary Activations

Compact ConvNets with Ternary Weights and Binary Activations

Compact architectures, ternary weights and binary activations are two methods suitable for making neural networks more efficient. We introduce a) a dithering binary activation which improves accuracy of ternary weight networks with binary activations by randomizing quantization error, and b) a metho...

Full description

Bibliographic Details
Main Author: Holesovsky, Ondrej
Format: Others
Language:English
Published: KTH, Robotik, perception och lärande, RPL 2017
Subjects:
convolutional neural networks
compact architectures
weight quantization
activation quantization
dithering activations
Computer Sciences
Datavetenskap (datalogi)
Online Access:http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-216389

Internet

http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-216389

Similar Items