02/02/2021

An Efficient Transformer Decoder with Compressed Sub-layers

Yanyang Li, Ye Lin, Tong Xiao, Jingbo Zhu

Keywords:

Abstract Paper Similar Papers

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.

The video of this talk cannot be embedded. You can watch it here:
https://slideslive.com/38948349
(Link will open in new window)
 0
 0
 0
 0
  Share    
This is an embedded video. Talk and the respective paper are published at AAAI 2021 virtual conference. If you are one of the authors of the paper and want to manage your upload, see the question "My papertalk has been externally embedded..." in the FAQ section.

Comments

Post Comment
no comments yet
code of conduct: tbd Characters remaining: 140

Similar Papers

 0
 0
 0
 0
 5:04 
 0
 0
 0
 0
 11:44 
 0
 0
 0
 0
 5:27 
 0
 1
 0
 1
 16:56 
 0
 0
 0
 0
 16:56 
 0
 0
 0
 0
 5:15 
 0
 0
 0
 0
 23:05 
 0
 0
 0
 0
 15:09 
 0
 0
 0
 0
 9:10 
 0
 0
 0
 0
 5:01 
 0
 0
 0
 0
 11:08 
 0
 0
 0
 0
 3:16 
 0
 0
 0
 0
 3:22 
 0
 0
 0
 0
 12:00 
 0
 0
 0
 0
 3:33 
 0
 0
 0
 0
 15:22 
 0
 0
 0
 0
 13:17 
 0
 0
 0
 0
 14:48 
 0
 0
 0
 0
 3:02 
 0
 0
 0
 0
 5:15 
 0
 0
 0
 0
 4:50 
 0
 0
 0
 1
 5:14 
 0
 0
 0
 0
 14:20 
 0
 0
 0
 0
 3:25 
 0
 0
 0
 0
 3:37 
 0
 0
 0
 0
 22:07 
 0
 0
 0
 0
 14:58 
 0
 0
 0
 0
 8:28 
 0
 0
 0
 0
 13:35 
 0
 0
 0
 0
 3:16 
 0
 0
 1
 0
 3:23 
 0
 0
 0
 0
 5:12 
 0
 0
 0
 0
 3:17 
 0
 0
 0
 0
 14:20 
 0
 0
 0
 0
 5:22 
 0
 0
 0
 0
 17:49 
 0
 0
 0
 0
 3:04 
 0
 0
 0
 0
 5:00 
 0
 0
 0
 0
 14:25 
 0
 0
 0
 0
 5:01