ResearchBenchmarksImage-to-Audio Retrieval on Unseen species-levelFollow28Top-1 AccuracyBioVITA28.7515.522.25Mar 25, 2026Evaluation ResultsMethodMethodLinksTop-1 AccuracyTop-5 AccuracyBioVITATraining Stage=Stage2Training Stage=Stage22026.032855.5BioVITATraining Stage=Stage1Training Stage=Stage12026.036.425.8ImageBind2026.03312.1