Fine-grained Visual Recognition on Bird-200 (test)
89.7EMSigLIP-FT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SigLIP-FTModel Category=CLIP-based2026.04 | 89.7 | 95.9 | |
| HyMOR-QwenModel Category=HyMOR framework2026.04 | 89 | 95.6 | |
| HyMOR-GemmaModel Category=HyMOR framework2026.04 | 88.8 | 95.5 | |
| BioCLIPModel Category=CLIP-based2026.04 | 88.7 | 95.3 | |
| Finedefics-8BModel Category=Fine-grained MLLMs2026.04 | 74.5 | 89.2 | |
| SigLIPModel Category=CLIP-based2026.04 | 51.9 | 79.8 | |
| Qwen2.5-VL-3BModel Category=General MLLMs2026.04 | 21.5 | 63.6 | |
| InternVL2.5-4BModel Category=General MLLMs2026.04 | 4.8 | 50.3 | |
| Gemma-3-4BModel Category=General MLLMs2026.04 | 2.1 | 53.7 |