VGGSound
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
VGGSound
1.51KL Divergence
4
VGGSound-AVEL 90K UCF to VGG
73.7Precision
4
VGGSound-AVEL 90K AVE to AVVP
60.2Segment-level F1
4
VGGSound-AVEL 90K AVE to AVVP
57.8Segment-level F1 Score
4
VGGSound-AVEL AVE to AVVP 40K
56.5Segment-level F1 Score
4
VGGSound All
38.8ALOHa
4
VGGSound 10 (test)
5.32FAD
4
VGGSOUND clean+
7.93Mean SDR
4
VGGSound V -> T
48Top-1 Accuracy
4
VGGSound A -> T
47.1Top-1 Accuracy
4
VGGSound
0FD
4
VGGSound
23.6Top-1 Accuracy
4
VGGSound (test)
138.8gFVD
3
VGGSound (fine-tuning)
48.91Accuracy
3
VGGSOUND clean+
15.91CLAPt Score
3
VGGSound
0.461LA-CLAP
3
VGGSound (test)
402.385FVD
3
VGGSound Paprika style (test)
2.612KL Divergence
3
VGGSound Multi-Caps
3.35Separability
2
VGGSound (test)
150.26gFVD
2
VGGSound All
40.52Cs
2
VGGSound
7SI-SDR
2
VGGSound
25.8Score
2
VGGSound Omini (off-screen track)
0.97FAD
2
VGGSound Mixed (test)
36PIAP
2