A New Storage-Less Hardware Compression Technique for Convolution Neural Networks Using Constant Multiplication. - Université Polytechnique des Hauts-de-France
Communication Dans Un Congrès Année : 2020

A New Storage-Less Hardware Compression Technique for Convolution Neural Networks Using Constant Multiplication.

Fichier non déposé

Dates et versions

hal-03389472 , version 1 (21-10-2021)

Identifiants

  • HAL Id : hal-03389472 , version 1

Citer

Mohamed Ayoub Neggaz, Ihsen Alouani, Smail Niar, Fadi Kurdahi. A New Storage-Less Hardware Compression Technique for Convolution Neural Networks Using Constant Multiplication.. WAICA: Workshop on Artificial Intelligence and Computer Architecture in Conjunction with HiPeac conference'20, Jan 2020, Bologne, Italy. ⟨hal-03389472⟩
46 Consultations
0 Téléchargements

Partager

More