ResearchBenchmarksCross-dataset domain transfer (Audio to Visual) on VGGSound-AVEL 90K AVE to AVVPCross-dataset domain transfer (Audio to Visual) on VGGSound-AVEL 90K AVE to AVVPFollow60.2Segment-level F1CoDAAR1.02416.38731.7547.113May 12, 2026Evaluation resultsMethodMethodLinksSegment-level F1CoDAARPretraining Scale=90KPretraining Scale=90K2026.0560.2DCIDPretraining Scale=90KPretraining Scale=90K2026.0557.5MICUPretraining Scale=90KPretraining Scale=90K2026.0551.6BaselinePretraining Scale=90KPretraining Scale=90K2026.053.3