Sharing attention weights in adjacent Transformer layers yields 1.3X inference speedup with negligible BLEU loss on ten WMT and NIST tasks.
Neural machine translation by jointly learning to align and translate
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Sharing Attention Weights for Fast Transformer
Sharing attention weights in adjacent Transformer layers yields 1.3X inference speedup with negligible BLEU loss on ten WMT and NIST tasks.