An Efficient Transformer Decoder with Compressed Sub-layers

Authors

  • Yanyang Li Northeastern University, China
  • Ye Lin Northeastern University, China
  • Tong Xiao Northeastern University, China
  • Jingbo Zhu Northeastern University, China

Keywords:

Machine Translation & Multilinguality

Abstract

The large attention-based encoder-decoder network (Transformer) has become prevailing recently due to its effectiveness. But the high computation complexity of its decoder raises the inefficiency issue. By examining the mathematic formulation of the decoder, we show that under some mild conditions, the architecture could be simplified by compressing its sub-layers, the basic building block of Transformer, and achieves a higher parallelism. We thereby propose Compressed Attention Network, whose decoder layer consists of only one sub-layer instead of three. Extensive experiments on 14 WMT machine translation tasks show that our model is 1.42x faster with performance on par with a strong baseline. This strong baseline is already 2x faster than the widely used standard baseline without loss in performance.

Downloads

Published

2021-05-18

How to Cite

Li, Y., Lin, Y., Xiao, T., & Zhu, J. (2021). An Efficient Transformer Decoder with Compressed Sub-layers. Proceedings of the AAAI Conference on Artificial Intelligence, 35(15), 13315-13323. Retrieved from https://ojs.aaai.org/index.php/AAAI/article/view/17572

Issue

Section

AAAI Technical Track on Speech and Natural Language Processing II