Binary ReRAM-based BNN first-layer implementation - CNRS - Centre national de la recherche scientifique Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Binary ReRAM-based BNN first-layer implementation

Résumé

The deployment of Edge AI requires energy-efficient hardware with a minimal memory footprint to achieve optimal performance. One approach to meet this challenge is the use of Binary Neural Networks (BNNs) based on non-volatile inmemory computing (IMC). In recent years, elegant ReRAM-based IMC solutions for BNNs have been developed, but they do not extend to the first layer of a BNN, which typically requires non-binary activations. In this paper, we propose a modified first layer architecture for BNNs that uses k-bit input images broken down into k binary input images with associated fully binary convolution layers and an accumulation layer with fixed weights of 2 −1 , ..., 2 −k. To further increase energy efficiency, we also propose reducing the number of operations by truncating 8-bit RGB pixel code to the 4 most significant bits (MSB). Our proposed architecture only reduces network accuracy by 0.28% on the CIFAR-10 task compared to a BNN baseline. Additionally, we propose a cost-effective solution to implement the weighted accumulation using successive charge sharing operations on an existing ReRAM-based IMC solution. This solution is validated through functional electrical simulations.
Fichier principal
Vignette du fichier
DATE23_VS.pdf (3.23 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04270562 , version 1 (04-11-2023)

Identifiants

Citer

Mona Ezzadeen, Atreya Majumdar, Sigrid Thomas, Jean-Philippe Noël, Bastien Giraud, et al.. Binary ReRAM-based BNN first-layer implementation. 2023 Design, Automation & Test in Europe Conference & Exhibition (DATE), Apr 2023, Antwerp, Belgium. pp.1-6, ⟨10.23919/DATE56975.2023.10137057⟩. ⟨hal-04270562⟩
41 Consultations
34 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More