Adapting Pretrained Text-to-Text Models for Long Text Sequences

Benchmark Model Rank Results
long-range-modeling-on-scrollsBART-LS#3Avg.: 39.76GovRep: 59.4 / 29.8 / 30.8SumScr: 37.7 / 10.2 / 21.5…
text-summarization-on-arxivBART-LS#2ROUGE-1: 50.2
text-summarization-on-booksumBART-LS#2ROUGE: 38.5
text-summarization-on-pubmedBART-LS#2ROUGE-1: 50.3