Fine-grained Species Generation on iNat mini (156 species)
30.36FIDTaxaAdapter
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| TaxaAdapterVTM Text Encoder=TaxaBind2026.03 | 30.36 | 0.741 | 57.82 | 81.47 | 30.97 | 24.86 | 0.36 | 0.86 | 0.3 | 0.92 | |
| TaxaAdapterVTM Text Encoder=BioCLIP2026.03 | 32.12 | 0.742 | 59.81 | 83.78 | 30.44 | 25.07 | 0.41 | 0.88 | 0.32 | 0.93 | |
| SD 3.5-LargeFine-tuning=LoRA2026.03 | 45.61 | 0.744 | 4.1 | 13.33 | 20.85 | 22.28 | 0.33 | 0.87 | 0.28 | 0.91 | |
| TaxaDiffusion2026.03 | 46.39 | 0.747 | — | — | 10.41 | — | — | — | — | — | |
| SD 1.5Mode=Zero-shot (ZS)2026.03 | 66.91 | 0.779 | 1.09 | 2.69 | 15.52 | 18.95 | 0.28 | 0.8 | 0.24 | 0.88 | |
| FLUX.1-devFine-tuning=LoRA2026.03 | 69.63 | 0.775 | 2.92 | 8.08 | 17.13 | 20.1 | 0.31 | 0.85 | 0.28 | 0.91 | |
| SD 3.5-LargeMode=Zero-shot (ZS)2026.03 | 70.98 | 0.802 | 2.12 | 7.05 | 16.19 | 19.36 | 0.31 | 0.81 | 0.26 | 0.9 | |
| FLUX.1-devMode=Zero-shot (ZS)2026.03 | 86.33 | 0.808 | 2.82 | 6.6 | 15.76 | 19.41 | 0.34 | 0.83 | 0.29 | 0.91 |