2 code implementations • ACL 2016 • Welin Chen, David Grangier, Michael Auli
Training neural network language models over large vocabularies is still computationally very costly compared to count-based models such as Kneser-Ney.
Machine Translation speech-recognition +2