Training high-performance and large-scale deep neural networks with full 8-bit integers

Deep neural network (DNN) quantization converting floating-point (FP) data in the network to integers (INT) is an effective way to shrink the model size for memory saving and simplify the operations for compute acceleration. Recently, researches on DNN quantization develop from inference to training...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	Neural networks 2020-05, Vol.125, p.70-82
Hauptverfasser:	Yang, Yukuan, Deng, Lei, Wu, Shuang, Yan, Tianyi, Xie, Yuan, Li, Guoqi
Format:	Artikel
Sprache:	eng
Schlagworte:	8-bit training Full quantization Neural network quantization Online learning device
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!