Fastformer

Reported on 6 benchmarks across 1 task · 1 paper

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

Knowledge Base6 results

Text SummarizationonPubmed
ROUGE-1· 2021-08-20
38.09
best: 51.05 (Top Down Transformer (AdaPool) (464M))
Fastformer: Additive Attention Can Be All You Need arXiv:2108.09084
Text SummarizationonPubmed
ROUGE-2· 2021-08-20
15.44
best: 24.76 (LongT5)
Fastformer: Additive Attention Can Be All You Need arXiv:2108.09084
Text SummarizationonPubmed
ROUGE-L· 2021-08-20
34.81
best: 46.67 (LongT5)
Fastformer: Additive Attention Can Be All You Need arXiv:2108.09084
Text SummarizationonCNN / Daily Mail (Anonymized)
ROUGE-1· 2021-08-20
38.54
best: 42.3 (HSSAS)
Fastformer: Additive Attention Can Be All You Need arXiv:2108.09084
Text SummarizationonCNN / Daily Mail (Anonymized)
ROUGE-2· 2021-08-20
16.22
best: 18.87 (RNES w/o coherence)
Fastformer: Additive Attention Can Be All You Need arXiv:2108.09084
Text SummarizationonCNN / Daily Mail (Anonymized)
ROUGE-L· 2021-08-20
36.21
best: 37.75 (RNES w/o coherence)
Fastformer: Additive Attention Can Be All You Need arXiv:2108.09084