Scaling Neural Machine Translation

Myle Ott, Sergey Edunov, David Grangier, Michael Auli


Abstract
Sequence to sequence learning models still require several days to reach state of the art performance on large benchmark datasets using a single machine. This paper shows that reduced precision and large batch training can speedup training by nearly 5x on a single 8-GPU machine with careful tuning and implementation. On WMT’14 English-German translation, we match the accuracy of Vaswani et al. (2017) in under 5 hours when training on 8 GPUs and we obtain a new state of the art of 29.3 BLEU after training for 85 minutes on 128 GPUs. We further improve these results to 29.8 BLEU by training on the much larger Paracrawl dataset. On the WMT’14 English-French task, we obtain a state-of-the-art BLEU of 43.2 in 8.5 hours on 128 GPUs.
Anthology ID:
W18-6301
Volume:
Proceedings of the Third Conference on Machine Translation: Research Papers
Month:
October
Year:
2018
Address:
Brussels, Belgium
Venues:
EMNLP | WMT | WS
SIG:
SIGMT
Publisher:
Association for Computational Linguistics
Note:
Pages:
1–9
Language:
URL:
https://www.aclweb.org/anthology/W18-6301
DOI:
10.18653/v1/W18-6301
Bib Export formats:
BibTeX MODS XML EndNote
PDF:
http://aclanthology.lst.uni-saarland.de/W18-6301.pdf