Mostra el registre d'ítem simple
Distributed training strategies for a computer vision deep learning algorithm on a distributed GPU cluster
dc.contributor.author | Campos, Victor |
dc.contributor.author | Sastre, Francesc |
dc.contributor.author | Yagües, Maurici |
dc.contributor.author | Bellver, Míriam |
dc.contributor.author | Giró Nieto, Xavier |
dc.contributor.author | Torres Viñals, Jordi |
dc.contributor.other | Universitat Politècnica de Catalunya. Departament de Teoria del Senyal i Comunicacions |
dc.contributor.other | Universitat Politècnica de Catalunya. Departament d'Arquitectura de Computadors |
dc.date.accessioned | 2017-09-13T10:24:11Z |
dc.date.available | 2017-09-13T10:24:11Z |
dc.date.issued | 2017 |
dc.identifier.citation | Campos, V., Sastre, F., Yagües, M., Bellver, M., Giro, X., Torres, J. Distributed training strategies for a computer vision deep learning algorithm on a distributed GPU cluster. "Procedia computer science", 2017, vol. 108, p. 315-324. |
dc.identifier.issn | 1877-0509 |
dc.identifier.uri | http://hdl.handle.net/2117/107590 |
dc.description.abstract | Deep learning algorithms base their success on building high learning capacity models with millions of parameters that are tuned in a data-driven fashion. These models are trained by processing millions of examples, so that the development of more accurate algorithms is usually limited by the throughput of the computing devices on which they are trained. In this work, we explore how the training of a state-of-the-art neural network for computer vision can be parallelized on a distributed GPU cluster. The effect of distributing the training process is addressed from two different points of view. First, the scalability of the task and its performance in the distributed setting are analyzed. Second, the impact of distributed training methods on the final accuracy of the models is studied. |
dc.format.extent | 10 p. |
dc.language.iso | eng |
dc.publisher | Elsevier |
dc.rights | Attribution-NonCommercial-NoDerivs 3.0 Spain |
dc.rights.uri | http://creativecommons.org/licenses/by-nc-nd/3.0/es/ |
dc.subject | Àrees temàtiques de la UPC::Enginyeria de la telecomunicació::Processament del senyal::Processament de la imatge i del senyal vídeo |
dc.subject | Àrees temàtiques de la UPC::Informàtica::Sistemes d'informació::Emmagatzematge i recuperació de la informació |
dc.subject.lcsh | Computer vision |
dc.subject.lcsh | Digital video |
dc.subject.lcsh | Neural networks (Computer science) |
dc.subject.lcsh | Graphics processing units |
dc.subject.lcsh | Computer algorithms |
dc.subject.other | Distributed computing |
dc.subject.other | Parallel systems |
dc.subject.other | Deep learning |
dc.subject.other | Convolutional Neural Networks |
dc.title | Distributed training strategies for a computer vision deep learning algorithm on a distributed GPU cluster |
dc.type | Article |
dc.subject.lemac | Visió per ordinador |
dc.subject.lemac | Vídeo digital |
dc.subject.lemac | Algorismes computacionals |
dc.subject.lemac | Xarxes neuronals (Informàtica) |
dc.subject.lemac | Unitats de processament gràfic |
dc.contributor.group | Universitat Politècnica de Catalunya. GPI - Grup de Processament d'Imatge i Vídeo |
dc.contributor.group | Universitat Politècnica de Catalunya. CAP - Grup de Computació d'Altes Prestacions |
dc.identifier.doi | 10.1016/j.procs.2017.05.074 |
dc.description.peerreviewed | Peer Reviewed |
dc.relation.publisherversion | http://www.sciencedirect.com/science/article/pii/S1877050917306129 |
dc.rights.access | Open Access |
local.identifier.drac | 21186189 |
dc.description.version | Postprint (published version) |
dc.relation.projectid | info:eu-repo/grantAgreement/MINECO//TEC2013-43935-R/ES/PROCESADO DE INFORMACION HETEROGENEA Y SEÑALES EN GRAFOS PARA BIG DATA. APLICACION EN CRIBADO DE ALTO RENDIMIENTO, TELEDETECCION, MULTIMEDIA Y HCI./ |
local.citation.author | Campos, V.; Sastre, F.; Yagües, M.; Bellver, M.; Giro, X.; Torres, J. |
local.citation.publicationName | Procedia computer science |
local.citation.volume | 108 |
local.citation.startingPage | 315 |
local.citation.endingPage | 324 |
Fitxers d'aquest items
Aquest ítem apareix a les col·leccions següents
-
Articles de revista [1.049]
-
Articles de revista [2.526]
-
Articles de revista [382]
-
Articles de revista [118]