Accelerating recurrent neural network training via two stage classes and parallelization.

Zhiheng Huang, Geoffrey Zweig, Michael Levit, Benot Dumoulin, Barlas Oguz, Shawn Chang

Browse the full ASRU paper archive.