HW-Flow: A Multi-Abstraction Level HW-CNN Codesign Pruning Methodology

HW-Flow: A Multi-Abstraction Level HW-CNN Codesign Pruning Methodology

Convolutional neural networks (CNNs) have produced unprecedented accuracy for many computer vision problems in the recent past. In power and compute-constrained embedded platforms, deploying modern CNNs can present many challenges. Most CNN architectures do not run in real-time due to the high numbe...

Full description

Bibliographic Details
Main Authors: Vemparala, Manoj-Rohit, Fasfous, Nael, Frickenstein, Alexander, Valpreda, Emanuele, Camalleri, Manfredi, Zhao, Qi, Unger, Christian, Nagaraja, Naveen-Shankar, Martina, Maurizio, Stechele, Walter
Format: Article
Language:English
Published: Schloss Dagstuhl -- Leibniz-Zentrum fuer Informatik 2022-11-01
Series:Leibniz Transactions on Embedded Systems
Subjects:
convolutional neural networks
optimization
hardware modeling
pruning
Online Access:https://drops.dagstuhl.de/storage/07lites/lites_vol008/lites_vol008_issue001/LITES.8.1.3/LITES.8.1.3.pdf

Internet

https://drops.dagstuhl.de/storage/07lites/lites_vol008/lites_vol008_issue001/LITES.8.1.3/LITES.8.1.3.pdf

Similar Items