HeadMask (Random-18)

Reported on 3 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 Processing3 results

Machine TranslationonIWSLT2015 Vietnamese-English
BLEU· 2020-09-21
26.85
SOTA
Alleviating the Inequality of Attention Heads for Neural Machine Translation arXiv:2009.09672
Machine TranslationonWMT2017 Turkish-English
BLEU score· 2020-09-21
17.56
SOTA
Alleviating the Inequality of Attention Heads for Neural Machine Translation arXiv:2009.09672
Machine TranslationonWMT2016 Romanian-English
BLEU score· 2020-09-21
32.85
best: 40.3 (fast-noisy-channel-modeling)
Alleviating the Inequality of Attention Heads for Neural Machine Translation arXiv:2009.09672