Joint image-text clustering on ImageNet 200 classes (val)
0.685GDOriginal CLIP
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Original CLIP2026.03 | 0.685 | 0.769 | 0.318 | |
| Mean-Centering2026.03 | 0.685 | 0.766 | 0.306 | |
| CLIP-Refine2026.03 | 0.647 | 0.765 | 0.31 | |
| M2-Mix2026.03 | 0.637 | 0.762 | 0.298 | |
| CS-Aligner2026.03 | 0.617 | 0.77 | 0.315 | |
| TPC-CMAalpha_target=0.012026.03 | 0.61 | 0.795 | 0.33 | |
| TPC-CMAalpha_target=0.052026.03 | 0.554 | 0.81 | 0.385 | |
| AlignCLIP2026.03 | 0.55 | 0.749 | 0.304 | |
| TPC-CMAalpha_target=0.32026.03 | 0.474 | 0.826 | 0.435 | |
| TPC-CMAalpha_target=0.52026.03 | 0.441 | 0.849 | 0.516 | |
| TPC-CMAalpha_target=0.72026.03 | 0.417 | 0.863 | 0.551 | |
| TPC-CMAalpha_target=0.92026.03 | 0.392 | 0.845 | 0.492 |