Context-dependent Text-to-SQL on CoSQL (dev)
58.8Question MatchRASAT + PICARD
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| RASAT + PICARDDecoding=PICARD2022.05 | 58.8 | 27 | 67 | 39.6 | — | — | |
| RASAT + PICARDApproach=Fine-tuned Model, Framework=PICARD2026.03 | 58.8 | 27 | 67 | 39.6 | — | — | |
| SFT Deepseek 7B + Track-SQLApproach=Fine-tuned, Backbone=Deepseek 7B, Framework=Track-SQL2026.03 | 58.19 | 28.67 | 70.6 | 43.67 | 62.26 | 32.76 | |
| SFT Mistral 7B + Track-SQLApproach=Fine-tuned, Backbone=Mistral 7B, Framework=Track-SQL2026.03 | 57.69 | 27.3 | 71.1 | 45.05 | 65.54 | 36.17 | |
| QDA-SQLApproach=Fine-tuned Model, Framework=QDA-SQL2026.03 | 57.3 | 30 | — | — | — | — | |
| T5-3B + PICARDDecoding=PICARD, Parameters=3B2022.05 | 56.9 | 24.2 | — | — | — | — | |
| HIE-SQL + GraPPaAugmentation=GraPPa2022.05 | 56.4 | 28.7 | — | — | — | — | |
| HIE-SQL + GraPPAApproach=Fine-tuned Model, Backbone=GraPPA2026.03 | 56.4 | 28.7 | — | — | — | — | |
| RASAT2022.05 | 56.2 | 25.9 | 63.8 | 34.8 | — | — | |
| SFT Mistral 7BApproach=Fine-tuned, Backbone=Mistral 7B2026.03 | 56.2 | 24.57 | 64.94 | 34.81 | 59.68 | 29.01 | |
| SFT Deepseek 7BApproach=Fine-tuned, Backbone=Deepseek 7B2026.03 | 54.71 | 23.2 | 66.03 | 34.12 | 58.88 | 26.96 | |
| UNIFIEDSKG2022.05 | 54.1 | 22.8 | 62.2 | 26.2 | — | — | |
| SFT Codellama 7B + Track-SQLApproach=Fine-tuned, Backbone=Codellama 7B, Framework=Track-SQL2026.03 | 53.82 | 24.57 | 66.73 | 37.88 | 59.18 | 29.01 | |
| T5-3BParameters=3B2022.05 | 53.8 | 21.8 | — | — | — | — | |
| CoE-SQLApproach=In-Context Learning, Backbone=GPT 3.52026.03 | 52.4 | 23.9 | 69.6 | 39.6 | 60.6 | 30.4 | |
| RAT-SQL + SCoReAugmentation=SCoRe2022.05 | 52.1 | 22 | — | — | — | — | |
| SFT Codellama 7BApproach=Fine-tuned, Backbone=Codellama 7B2026.03 | 50.54 | 17.74 | 60.17 | 28.66 | 54.51 | 23.54 | |
| ACT-SQLApproach=In-Context Learning, Backbone=GPT 3.52026.03 | 46 | 13.3 | 63.7 | 30.7 | 55.2 | 21.5 | |
| Dynamic Hybrid Relation NetworkPre-trained Model=BERT2021.01 | 45.7 | 19.5 | — | — | — | — | |
| GAZP + BERTBackbone=BERT2022.05 | 42 | 12.3 | 38.8 | — | — | — | |
| RichContextPre-trained Model=BERT2021.01 | 41 | 14 | — | — | — | — | |
| EditSQLPre-trained Model=BERT2021.01 | 39.9 | 12.3 | — | — | — | — | |
| Dynamic Hybrid Relation NetworkPre-trained Model=GloVe2021.01 | 34.5 | 11 | — | — | — | — | |
| RichContextPre-trained Model=GloVe2021.01 | 33.5 | 9.6 | — | — | — | — | |
| Liu et al. (2020)Enc-Dec=HE-Grammar2021.09 | 33.5 | 9.6 | — | — | — | — | |
| MemCE+Liu et al. (2020)Enc-Dec=Mem-Grammar2021.09 | 32.8 | 10.6 | — | — | — | — | |
| EditSQLPre-trained Model=GloVe2021.01 | 22.2 | 5.8 | — | — | — | — | |
| SyntaxSQL-conPre-trained Model=GloVe2021.01 | 15.1 | 2.7 | — | — | — | — | |
| CD-Seq2SeqPre-trained Model=GloVe2021.01 | 13.8 | 2.1 | — | — | — | — | |
| CDS2SEnc-Dec=HE-LSTM2021.09 | 13.8 | 2.1 | — | — | — | — | |
| MemCE+CDS2SEnc-Dec=Mem-LSTM2021.09 | 13.4 | 3.4 | — | — | — | — | |
| MemCE+CDS2SEnc-Dec=Mem-SnipCopy2021.09 | 13.1 | 2.7 | — | — | — | — | |
| CDS2SEnc-Dec=HE-SnipCopy2021.09 | 12.3 | 2.1 | — | — | — | — |