"A scalable GPU-enabled framework for training deep neural networks."

Bonaventura Del Monte, Radu Prodan (2016)
a service of Schloss Dagstuhl - Leibniz Center for Informatics