Low-Precision Mixed-Computation Models for Inference on Edge

This article presents a mixed-computation neural network processing approach for edge applications that incorporates low-precision (low-width) Posit and low-precision fixed point (FixP) number systems. This mixed-computation approach uses 4-bit Posit (Posit4), which has higher precision around 0, fo...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Veröffentlicht in:	IEEE transactions on very large scale integration (VLSI) systems 2024-08, Vol.32 (8), p.1414-1422
Hauptverfasser:	Azizi, Seyedarmin, Nazemi, Mahdi, Kamal, Mehdi, Pedram, Massoud
Format:	Artikel
Sprache:	eng
Schlagworte:	Accuracy Back propagation networks Computation Computational modeling Decoding Deep neural network (DNN) model energy Energy consumption Error analysis fixed-point Hardware low-precision mixed-computation Neural networks Number systems posit Quantization (signal) Training
Online-Zugang:	Volltext bestellen
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Schreiben Sie den ersten Kommentar!