Reference : CalcGraph: taming the high costs of deep learning using models

	Document type :	Scientific journals : Article
	Discipline(s) :	Engineering, computing & technology : Computer science
	Focus Areas :	Computational Sciences
	To cite this reference:	http://hdl.handle.net/10993/52860

Title :	CalcGraph: taming the high costs of deep learning using models
Language :	English
Author, co-author :	Lorentz, Joe [DataThings S.A. > > > ; University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT)]
	Hartmann, Thomas [DataThings S.A.]
	Moawad, Assaad [DataThings S.A.]
	Fouquet, François [DataThings S.A.]
	Aouada, Djamila [University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > CVI2 >]
	Le Traon, Yves [University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > SerVal >]
Publication date :	25-Oct-2022
Journal title :	Software and Systems Modeling
Publisher :	Springer
Peer reviewed :	Yes
ISSN :	1619-1366
e-ISSN :	1619-1374
Country :	Germany
Keywords :	[en] Differentiable programming ; Computational graph model ; Edge AI
Abstract :	[en] Models based on differential programming, like deep neural networks, are well established in research and able to outperform manually coded counterparts in many applications. Today, there is a rising interest to introduce this flexible modeling to solve real-world problems. A major challenge when moving from research to application is the strict constraints on computational resources (memory and time). It is difficult to determine and contain the resource requirements of differential models, especially during the early training and hyperparameter exploration stages. In this article, we address this challenge by introducing CalcGraph, a model abstraction of differentiable programming layers. CalcGraph allows to model the computational resources that should be used and then CalcGraph’s model interpreter can automatically schedule the execution respecting the specifications made. We propose a novel way to efficiently switch models from storage to preallocated memory zones and vice versa to maximize the number of model executions given the available resources. We demonstrate the efficiency of our approach by showing that it consumes less resources than state-of-the-art frameworks like TensorFlow and PyTorch for single-model and multi-model execution.
Permalink :	http://hdl.handle.net/10993/52860
Mentions required by the publisher for OA :	This version of the article has been accepted for publication, after peer review but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1007/s10270-022-01052-7
FnR project :	FnR ; FNR14297122 > Joe Lorentz > DETECT > Towards Edge-optimized Deep Learning For Explainable Quality Control > 01/01/2020 > 31/12/2023 > 2019

	File(s) associated to this reference
	Fulltext file(s): File Commentary Version Size Access Open access accepted_version_CalcGraph: taming the high costs of deep learning using models.pdf Author postprint 740.99 kB View/Open

All documents in ORBi^lu are protected by a user license.

 

University of Luxembourg Library |  Feedback |  Legal notices Site Map