AudioCaps
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
AudioCaps 1K 1.0 (test)
52Recall@1
10
AudioCaps AudioSet (test)
48.5SPIDEr
10
AudioCaps
0.09BER (Original)
8
AudioCaps (test)
96.926FAD
8
AudioCaps 1K 1.0 (test)
52.4R@1
8
AudioCaps (test)
8.59IS
7
AudioCaps (test)
0.457SRCC
7
AudioCaps 5 hypotheses (test)
0.735Div2 Score
6
AudioCaps T+V -> A
95.2Recall@1
6
AudioCaps (val)
3.2VISQOL Score
6
AudioCaps Subset
1.883LSD
5
AudioCaps Subset
1.832LSD
5
AudioCaps Subset
1.795LSD
5
AudioCaps public (test)
0.032Mel Distance
5
AudioCaps multi-event prompts
75.2FDopenl3
5
AudioCaps (test)
13.28IS (Inception Score)
5
AudioCaps (val)
21.98mAP
5
AudioCaps (test)
9.76FID
5
AudioCaps 3-source mixture (test)
13.3SI-SNRi
4
AudioCaps 2-source mixture (test)
20.3SI-SNRi
4
AudioCaps (val)
64CIDEr
4
AudioCaps-QA
54.31FENSE Score
3
AudioCaps InD, 10s (test)
2.01FAD
3
AudioCaps 3-source mixture (test)
77.8Accuracy
3
AudioCaps 2-source mixture (test)
89.9Accuracy
3