Clustering on Caltech-101
0.9NMIDIVA
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| DIVABackbone=SigLIP ViT-SO@224 (S-1), Zero-shot=true2026.03 | 0.9 | 66 | 0.48 | |
| Original CLIPBackbone=SigLIP ViT-SO@224 (S-1), Zero-shot=true2026.03 | 0.89 | 61 | 0.43 | |
| DCRBackbone=SigLIP ViT-SO@224 (S-1), Zero-shot=true2026.03 | 0.89 | 66 | 0.48 | |
| GenHancerBackbone=SigLIP ViT-SO@224 (S-1), Zero-shot=true2026.03 | 0.88 | 64 | 0.48 | |
| DCRBackbone=OpenAI CLIP ViT-L@224 (O-1), Zero-shot=true2026.03 | 0.83 | 57 | 0.43 | |
| DCRBackbone=MetaCLIP ViT-L@224 (M-1), Zero-shot=true2026.03 | 0.82 | 57 | 0.42 | |
| GenHancerBackbone=OpenAI CLIP ViT-L@224 (O-1), Zero-shot=true2026.03 | 0.81 | 54 | 0.38 | |
| GenHancerBackbone=MetaCLIP ViT-L@224 (M-1), Zero-shot=true2026.03 | 0.81 | 55 | 0.41 | |
| un2CLIPBackbone=OpenAI CLIP ViT-L@224 (O-1), Zero-shot=true2026.03 | 0.8 | 55 | 0.41 | |
| DIVABackbone=MetaCLIP ViT-L@224 (M-1), Zero-shot=true2026.03 | 0.8 | 56 | 0.39 | |
| Original CLIPBackbone=MetaCLIP ViT-L@224 (M-1), Zero-shot=true2026.03 | 0.79 | 52 | 0.4 | |
| Original CLIPBackbone=OpenAI CLIP ViT-L@224 (O-1), Zero-shot=true2026.03 | 0.78 | 53 | 0.35 | |
| DIVABackbone=OpenAI CLIP ViT-L@224 (O-1), Zero-shot=true2026.03 | 0.78 | 52 | 0.34 |