Text-conditioned 3D generation on ShapeNetCore13 (test)
88.04AccuracyVPP
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| VPP2023.07 | 88.04 | 29.82 | 10.64 | |
| CLIP-Sculptor2023.07 | 87.08 | 1,480.11 | — | |
| CLIP-Forge2023.07 | 83.33 | 2,425.25 | — | |
| Point-Evariant=40M text-only, protocol=zero-shot transfer2023.07 | 62.56 | 85.37 | 7.93 |