Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

BIRD

Benchmarks

Task NameDataset NameSOTA ResultTrend
Text-to-SQLBIRD (dev)
Execution Accuracy (EA)74.46
251
Text-to-SQLBIRD
Total Execution Accuracy70.5
64
Text-to-SQLBIRD (test)
EX75.63
46
Text2SQLBIRD (dev)
Exec Acc (Greedy)65.8
44
Image ClassificationBird
Accuracy85.1
29
Color Video CompletionBird color video sequence
PSNR25.382
28
Text-to-SQLBIRD
Accuracy69.1
27
SQL Semantic ValidationBIRD
AUPRC80.36
24
SQL execution performanceBIRD n=1534
EM (1T)64.9
21
Schema retrievalBIRD
Recall (R)100
15
SQL GenerationBIRD Original (dev)
Execution Accuracy (Simple)65.51
14
SQL GenerationBIRD Verified
Execution Accuracy (Simple)69.41
14
Text-to-SQLBIRD (test dev)
Execution Accuracy (EX)48.92
14
Table SelectionBIRD 2023 (test)
Avg #tables5.3
12
Text-to-SQLBIRD (holdout test)
Execution Accuracy73
11
Text2SQLBIRD Movies
Execution Accuracy46.9
9
Text2SQLBIRD App Store
Execution Accuracy38.4
9
Text2SQLBIRD Computer Students
Execution Accuracy48.3
9
Table RetrievalBIRD
Capped Recall@2599.1
9
SQL generationBird
Greedy70.5
9
RetrievalBird
Recall@1096.1
9
Text-to-SQLBIRD Movies
Spearman Correlation0.65
8
Text-to-SQLBIRD Apps
Spearman Correlation0.74
8
Text-to-SQLBIRD Computer
Spearman Correlation0.46
8
Table RetrievalBird (dev)
Precision@586.2
8
Showing 25 of 44 rows