Video-to-Audio Retrieval on MSR-VTT
17.2Recall@1AVLnet
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| AVLnetZero-shot=true, Pre-train=HowTo100M2024.03 | 17.2 | 26.6 | 46.6 | |
| VAB-EncodecDirection=Video -> Audio, Zero-shot=true2024.09 | 14.2 | 31.1 | 42 | |
| EquiAVZero-shot=true2024.03 | 13.8 | 31.4 | 43 | |
| VAB-DACDirection=Video -> Audio, Zero-shot=true2024.09 | 13.8 | 30.6 | 40.1 | |
| CAV-MAEZero-shot=true2024.03 | 13.3 | 29 | 40.5 | |
| CAV-MAEDirection=Video -> Audio, Zero-shot=true2024.09 | 13.3 | 29 | 40.5 | |
| AranjelovicZero-shot=true, Pre-train=HowTo100M2024.03 | 11.9 | 25.9 | 34.7 | |
| BoggustZero-shot=true, Pre-train=HowTo100M2024.03 | 9.3 | 20.7 | 28.8 | |
| ImageBindDirection=Video -> Audio, Zero-shot=true2024.09 | 7 | 18.5 | 25.2 | |
| LanguageBindDirection=Video -> Audio, Zero-shot=true2024.09 | 1.9 | 6.2 | 8.8 |