Text-to-Audio Retrieval on AudioCaps 1K 1.0 (test)
52Recall@1VAST
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| VAST# (audio, text)=27M, Training Regime=internet-scale2023.12 | 52 | 76.8 | 82.9 | |
| FuseMix(W&H,B)# (audio, text)=65K, Training Regime=low-data regime2023.12 | 43.1 | 77.4 | 87.4 | |
| ONE-PEACE# (audio, text)=2.4M, Training Regime=internet-scale2023.12 | 42.5 | 77.5 | 88.4 | |
| FuseMix(W&H,B)# (audio, text)=50K/15K, Training Regime=low-data regime2023.12 | 41.3 | 76.9 | 87.6 | |
| VALOR# (audio, text)=6.5M, Training Regime=internet-scale2023.12 | 40.1 | 73.9 | 83.1 | |
| LAION-CLAP# (audio, text)=65K, Training Regime=low-data regime2023.12 | 36.7 | 70.9 | 83.2 | |
| LAION-CLAP# (audio, text)=700K, Training Regime=internet-scale2023.12 | 36.2 | 70.3 | 82.5 | |
| MMT# (audio, text)=50K/15K, Training Regime=low-data regime2023.12 | 36.1 | 72 | 84.5 | |
| CLAP-HTSAT# (audio, text)=70K, Training Regime=low-data regime2023.12 | 34.7 | 70.2 | 82 | |
| ML-ACT# (audio, text)=50K/15K, Training Regime=low-data regime2023.12 | 33.9 | 69.7 | 82.6 |