XLNet+DSC

Reported on 4 benchmarks across 1 task · 1 paper · 1 SOTA

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

Natural Language Processing4 results

Question AnsweringonSQuAD1.1 dev
F1· uses extra data· 2019-11-07
95.77
SOTA
Dice Loss for Data-imbalanced NLP Tasks arXiv:1911.02855
Question AnsweringonSQuAD1.1 dev
EM· uses extra data· 2019-11-07
89.79
best: 90.06 (T5-11B)
Dice Loss for Data-imbalanced NLP Tasks arXiv:1911.02855
Question AnsweringonSQuAD2.0 dev
EM· 2019-11-07
87.65
best: 87.9 (XLNet (single model))
Dice Loss for Data-imbalanced NLP Tasks arXiv:1911.02855
Question AnsweringonSQuAD2.0 dev
F1· 2019-11-07
89.51
best: 90.6 (XLNet (single model))
Dice Loss for Data-imbalanced NLP Tasks arXiv:1911.02855