VGGSound
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
VGGSound (test)
0.52FAD
95
VGGSound single-source (test)
60.2IoU@0.5
39
VGGSound GZSL (test)
29.96S Score
38
VGGSound-Duet (test)
77.6CIoU@0.3
37
VGGSound
69.8Top-1 Acc
37
VGGSound
0.97FD_VGG
32
VGGSound-C unimodal (test)
53.14Accuracy (Gaussian)
25
VGGSound (test)
35.41IB-score
25
VGGSound-C (test)
6.2Error Rate (Gauss.)
24
VGGSound (test)
69.1Fusion Top-1 Acc
23
VGGSound-C severity level 5 (test)
54.9Gauss. Corruption Accuracy
20
VGGSound-S (test)
6.8Recall@1 (Video -> Text)
19
VGGSound-AVEL 90K
67Precision
15
VGGSound-AVEL 40K
75.3Precision
15
VGGSound
33.5R@1
15
VGGSound Animal
51.52Cs Score
14
VGGSound-C severity level 5
54.7Accuracy (Gaussian Blur)
14
VGGSound
52.7Zero-shot Accuracy
14
VGGSound GZSL
48.33S Score
14
VGGSound-2C bimodal (test)
43.74Accuracy (Gaussian)
14
VGGSound
1.44Fréchet Audio Distance (FAD)
14
VGGSound (test)
56.78Accuracy
13
VGGSound (test)
34.9Recall@1
13
VGGSound Instruments (test)
89.6CIoU@0.1
13
VGGSound Instruments (test)
69.5IoU@0.3
13