Audio-to-Video Retrieval on VGGSound (test)
34.9Recall@1VAB-Encodec
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| VAB-EncodecEvaluation Protocol=Fine-tuned2026.04 | 34.9 | 62.7 | 73.1 | |
| GMS-CAVP2026.01 | 30.5 | 45.3 | 58.2 | |
| TG-DPEvaluation Protocol=Zero-shot2026.04 | 30.3 | 53.1 | 64.2 | |
| DenseAVEvaluation Protocol=Zero-shot2026.04 | 25.1 | 50.1 | 59.1 | |
| WAVE 7B2025.09 | 25 | — | — | |
| CAV-MAE SyncEvaluation Protocol=Zero-shot2026.04 | 23.2 | 46.2 | 58.1 | |
| ImageBindEvaluation Protocol=Zero-shot2026.04 | 20.7 | 43.2 | 53.4 | |
| AVSiamEvaluation Protocol=Zero-shot2026.04 | 20.4 | — | — | |
| CAV-MAE Scale+++Evaluation Protocol=Zero-shot2026.04 | 12.8 | 30.4 | 40.3 | |
| CAVP2026.01 | 11.1 | 27.8 | 36.4 | |
| encoder-only retrieval modelReference Type=Reference Model2025.09 | 10.3 | — | — | |
| CAV-MAEEvaluation Protocol=Zero-shot2026.04 | 8.3 | 23.8 | 32.4 | |
| LanguageBindEvaluation Protocol=Zero-shot2026.04 | 6.5 | 22.7 | 33.5 |