Audio-to-Text Retrieval on VGGSounder
48.9mAPAuroLA
Evaluation results
| Method | Links | |
|---|---|---|
| AuroLAPre-training setting=Full AudioVerse training2026.02 | 48.9 | |
| AuroLAPre-training setting=PT2026.02 | 46.6 | |
| PE-AVPre-training setting=PT2026.02 | 45.8 | |
| WavCapsPre-training setting=Standard2026.02 | 44.6 | |
| OnePeacePre-training setting=PT2026.02 | 42.3 | |
| VASTPre-training setting=PT2026.02 | 36.2 | |
| CLAP (fusion)Pre-training setting=Standard, Model Variant=Fusion2026.02 | 32.4 | |
| ReCLAPPre-training setting=Standard2026.02 | 17 |