Video-to-Audio Retrieval on VGGSound (test)
33.5Recall@1VAB-Encodec
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| VAB-EncodecEvaluation Protocol=Fine-tuned2026.04 | 33.5 | 63.3 | 74.3 | |
| TG-DPEvaluation Protocol=Zero-shot2026.04 | 31.3 | 54.4 | 63.9 | |
| CAV-MAE SyncEvaluation Protocol=Zero-shot2026.04 | 27.9 | 51.7 | 61.8 | |
| DenseAVEvaluation Protocol=Zero-shot2026.04 | 24.2 | 48.2 | 58.3 | |
| ImageBindEvaluation Protocol=Zero-shot2026.04 | 21.6 | 43.4 | 52.9 | |
| AVSiamEvaluation Protocol=Zero-shot2026.04 | 19 | — | — | |
| CAV-MAE Scale+++Evaluation Protocol=Zero-shot2026.04 | 14.8 | 34.2 | 44 | |
| CAV-MAEEvaluation Protocol=Zero-shot2026.04 | 14.7 | 35.3 | 45.9 | |
| LanguageBindEvaluation Protocol=Zero-shot2026.04 | 10.3 | 30.1 | 39.7 | |
| GMS-CAVP2026.01 | 0.289 | 0.437 | 0.579 | |
| CAVP2026.01 | 0.095 | 0.254 | 0.351 |