UNIQ: Uniform Noise Injection for Non-Uniform Quantization of Neural Networks

We present a novel method for neural network quantization. Our method, named UNIQ , emulates a non-uniform k -quantile quantizer and adapts the model to perform well with quantized weights by injecting noise to the weights at training time. As a by-product of injecting noise to weights, we find that...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	ACM transactions on computer systems 2021-06, Vol.37 (1-4), p.1-15
Hauptverfasser:	Baskin, Chaim, Liss, Natan, Schwartz, Eli, Zheltonozhskii, Evgenii, Giryes, Raja, Bronstein, Alex M., Mendelson, Avi
Format:	Artikel
Sprache:	eng
Schlagworte:	Accuracy Complexity Measurement Neural networks Noise
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!