Table Question Answering on WIKISQL WEAK (test)
90.4Denotation AccuracyREASTAP
Evaluation results
| Method | Links | |
|---|---|---|
| REASTAPbackbone=BART-Large2022.10 | 90.4 | |
| TAPEX2021.07 | 89.5 | |
| TAPEX2022.10 | 89.2 | |
| Min et al. (2019)Execution-Guided Decoding=true2021.07 | 87.2 | |
| BARTbackbone=BART-Large2022.10 | 86.1 | |
| T5-3Bsize=3B2022.10 | 86 | |
| BART2021.07 | 85.8 | |
| Yu et al. (2021a)2021.07 | 84.7 | |
| GraPPa2022.10 | 84.7 | |
| Min et al. (2019)2021.07 | 83.9 | |
| HardEM2022.10 | 83.9 | |
| Herzig et al. (2020)2021.07 | 83.6 | |
| TaPas2022.10 | 83.6 | |
| Wang et al. (2019b)2021.07 | 79.3 | |
| LatentAlignment2022.10 | 79.3 | |
| Agarwal et al. (2019)2021.07 | 74.8 | |
| MeRL2022.10 | 74.8 | |
| Liang et al. (2018)2021.07 | 72.4 | |
| MAPO2022.10 | 72.4 | |
| Guo & Gao (2019)2021.07 | 61 |