Transformer+WDrop

Reported on 3 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.

Knowledge Base3 results

Text SummarizationonDUC 2004 Task 1
ROUGE-1· uses extra data· 2021-04-05
33.06
SOTA
Rethinking Perturbations in Encoder-Decoders for Fast Training arXiv:2104.01853
Text SummarizationonDUC 2004 Task 1
ROUGE-2· uses extra data· 2021-04-05
11.45
best: 17.74 (Transformer LM)
Rethinking Perturbations in Encoder-Decoders for Fast Training arXiv:2104.01853
Text SummarizationonDUC 2004 Task 1
ROUGE-L· uses extra data· 2021-04-05
28.51
best: 28.52 (Transformer+LRPE+PE+Re-ranking+Ensemble)
Rethinking Perturbations in Encoder-Decoders for Fast Training arXiv:2104.01853