ResearchBenchmarksCross-modal Retrieval (Image to Text) on iNaturalist Species-level Retrieval (seen)Cross-modal Retrieval (Image to Text) on iNaturalist Species-level Retrieval (seen)Follow86.3Top-1 AccuracyBioVITA-3.441619.856743.15566.4533Mar 25, 2026Evaluation resultsMethodMethodLinksTop-1 AccuracyTop-5 AccuracyBioVITATraining Stage=Stage2Training Stage=Stage22026.0386.396BioCLIP 22026.0365.180.5BioVITATraining Stage=Stage1Training Stage=Stage12026.0365.180.5ImageBind2026.0359.581.4TaxaBind2026.0356.980CLIP2026.0350.965.1Random2026.030.010.05