ResearchBenchmarksAudio-to-Image Retrieval on Unseen species-levelAudio-to-Image Retrieval on Unseen species-levelFollow29.3Top-1 AccuracyBioVITA-0.0287.58615.222.814Mar 25, 2026Evaluation resultsMethodMethodLinksTop-1 AccuracyTop-5 AccuracyBioVITATraining Stage=Stage2Training Stage=Stage22026.0329.354.3BioVITATraining Stage=Stage1Training Stage=Stage12026.037.823.6ImageBind2026.031.19.2