Transformer (big) + Relative Position Representations

Reported on 2 benchmarks across 1 task · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing2 results

Machine TranslationonWMT2014 English-German
BLEU score· 2018-03-06
29.2
best: 35.14 (Transformer Cycle (Rev))
SOTA
Self-Attention with Relative Position Representations arXiv:1803.02155
Machine TranslationonWMT2014 English-French
BLEU score· 2018-03-06
41.5
best: 46.4 (Transformer+BT (ADMIN init))
SOTA
Self-Attention with Relative Position Representations arXiv:1803.02155