Communication Dans Un Congrès Année : 2025

Hardware-Aware Training for Multiplierless Convolutional Neural Networks

Résumé

Many computer vision tasks use convolutional neural networks (CNNs). These networks have a significant computational cost and complex implementations, in particular on embedded systems. A common way to implement CNNs on integrated circuits is to use low-precision quantized weights and activations instead of de facto floating-point (FP) ones. This is important to reduce the implementation cost. However, this has a drawback regarding accuracy, and Quantization-Aware Training (QAT) is one of the most popular approaches to mitigate this issue. In this article, we introduce an multiplierlessaware training approach that significantly reduces hardware resource consumption. We propose to incrementally fix weights to their current value based on their implementation cost. To compute this cost, we base our approach on Multiple Constant Multiplication (MCM) shift-and-add solving technique. With this idea, we show a global implementation cost reduction by around 25% w. r. t. a vanilla QAT approach without hardware usage in the loop. Compared to state-of-the-art multiplierless-aware training methods, the network accuracy of our designs is closer to that of a vanilla QAT baseline.
Fichier principal
Vignette du fichier
Shift_and_Add_NN.pdf (695) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04949886 , version 1 (15-02-2025)

Licence

Identifiants

  • HAL Id : hal-04949886 , version 1

Citer

Rémi Garcia, Léo Pradels, Silviu-Ioan Filip, Olivier Sentieys. Hardware-Aware Training for Multiplierless Convolutional Neural Networks. ARITH 2025 - 32nd IEEE International Symposium on Computer Arithmetic, May 2025, El Paso, United States. pp.1-8. ⟨hal-04949886⟩
0 Consultations
0 Téléchargements

Partager

More