Text-to-SQL on WikiSQL Fully-supervised (test)
90.8Execution AccuracyGuo & Gao w. GRAPPA (MLM+SSP)
Evaluation Results
| Method | Links | |
|---|---|---|
| Guo & Gao w. GRAPPA (MLM+SSP)Execution-guided decoding=false, Backbone=GRAPPA, Pre-training Objective=MLM+SSP2020.09 | 90.8 | |
| Guo & Gao w. GRAPPA (MLM)Execution-guided decoding=false, Backbone=GRAPPA, Pre-training Objective=MLM2020.09 | 90.7 | |
| Guo & Gao w. GRAPPA (SSP)Execution-guided decoding=false, Backbone=GRAPPA, Pre-training Objective=SSP2020.09 | 90.7 | |
| Guo & Gao w. RoBERTa-largeExecution-guided decoding=false, Backbone=RoBERTa-large2020.09 | 90.6 | |
| Lyu et al.Execution-guided decoding=false2020.09 | 89.2 | |
| Guo & GaoExecution-guided decoding=false2020.09 | 89.2 | |
| He et al.Execution-guided decoding=false2020.09 | 88.7 | |
| Hwang et al.Execution-guided decoding=false2020.09 | 86.2 | |
| Shi et al.Execution-guided decoding=false2020.09 | 83.7 | |
| Guo & Gao w. GRAPPA (MLM+SSP) (10k)Execution-guided decoding=false, Backbone=GRAPPA, Pre-training Objective=MLM+SSP, Training Data Scale=10k2020.09 | 82.2 | |
| Guo & Gao w. RoBERTa-large (10k)Execution-guided decoding=false, Backbone=RoBERTa-large, Training Data Scale=10k2020.09 | 79.2 | |
| Dong & LapataExecution-guided decoding=false2020.09 | 78.5 |