Base-to-new class generalization on ModelNet40
97.17Base AccuracyReFine3D
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ReFine3DPre-trained Model=ULIP2026.06 | 97.17 | 58.3 | 72.88 | |
| ReFine3DPre-trained Model=ULIP-22026.06 | 95.87 | 66.23 | 78.34 | |
| PointPRCPre-trained Model=ULIP-22026.06 | 95.3 | 64.83 | 77.17 | |
| PointPRCPre-trained Model=ULIP2026.06 | 95.03 | 55.27 | 69.89 | |
| P-CLIP2Pre-trained Model=ULIP2026.06 | 93.98 | 45.21 | 61.05 | |
| P-CLIPPre-trained Model=ULIP2026.06 | 93.23 | 20.22 | 33.23 | |
| ULIPPre-trained Model=ULIP2026.06 | 92.8 | 50.07 | 65.05 | |
| ULIP-2Pre-trained Model=ULIP-22026.06 | 91.77 | 56.47 | 69.92 |