Evaluating Mixed-Precision Arithmetic for 3D Generative Adversarial Networks to Simulate High Energy Physics Detectors

Several hardware companies are proposing native Brain Float 16-bit (BF16) support for neural network training. The usage of Mixed Precision (MP) arithmetic with floating-point 32-bit (FP32) and 16-bit half-precision aims at improving memory and floating-point operations throughput, allowing faster t...

Descripción completa

Detalles Bibliográficos
Autores principales: Ríos, John Osorio, Armejach, Adrià, Khattak, Gulrukh, Petit, Eric, Vallecorsa, Sofia, Casas, Marc
Lenguaje:eng
Publicado: 2020
Acceso en línea:https://dx.doi.org/10.1109/ICMLA51294.2020.00017
http://cds.cern.ch/record/2759602

Ejemplares similares

Ejemplares similares