Abstractive Summarization on CNNDM (test)
47.78ROUGE-1BRIO-Mul
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| BRIO-MulBackbone=BART, Training Loss=Multi-task Loss, Training Mode=Fine-tuned2022.03 | 47.78 | 23.55 | 44.57 | |
| BRIO-CtrBackbone=BART, Training Loss=Contrastive Loss Only, Training Mode=Fine-tuned2022.03 | 47.28 | 22.93 | 44.15 | |
| SimCLSBackbone=BART, Evaluation Model=RoBERTa, Evaluation Source=Original Paper2022.03 | 46.67 | 22.15 | 43.54 | |
| GSumBackbone=BART, Evaluation Source=Original Paper2022.03 | 45.94 | 22.32 | 42.48 | |
| GOLD-pBackbone=BART, Evaluation Source=Original Paper2022.03 | 45.4 | 22.01 | 42.25 | |
| SeqCoBackbone=BART, Evaluation Source=Original Paper2022.03 | 45.02 | 21.8 | 41.75 | |
| GOLD-sBackbone=BART, Evaluation Source=Original Paper2022.03 | 44.82 | 22.09 | 41.81 | |
| ConSumBackbone=BART, Evaluation Source=Original Paper2022.03 | 44.53 | 21.54 | 41.57 | |
| BARTBackbone=BART, Evaluation Source=Current Paper Evaluation Script2022.03 | 44.29 | 21.17 | 41.09 | |
| PEGASUSEvaluation Source=Original Paper2022.03 | 44.17 | 21.47 | 41.11 | |
| BARTBackbone=BART, Evaluation Source=Original Paper2022.03 | 44.16 | 21.28 | 40.9 |