3D Mesh Captioning on PointLLM-200 Objaverse
65.91GPT Score (Image-based)EVA01
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| EVA01variant=Final2026.05 | 65.91 | 6.334 | 0.585 | 9.323 | 13.566 | 50.554 | 50.878 | 59.095 | |
| EVA01variant=Align2026.05 | 56.77 | 23.592 | 3.555 | 28.002 | 19.007 | 52.178 | 52.005 | 48.825 | |
| GT caption2026.05 | 56.05 | — | — | — | — | — | — | — | |
| PointLLMParameters=7B2026.05 | 51.495 | 7.905 | 0.669 | 10.29 | 13.536 | 47.894 | 48.812 | 52.36 | |
| GreenPLM2026.05 | 50.475 | 17.527 | 2.303 | 20.632 | 19.906 | 47.826 | 48.871 | 50.305 | |
| PointLLMParameters=13B2026.05 | 49.745 | 7.873 | 0.649 | 10.519 | 13.62 | 47.539 | 48.602 | 51.735 | |
| ShapeLLMParameters=13B2026.05 | 35.87 | 10.542 | 1.05 | 12.954 | 14.234 | 39.935 | 40.728 | 33.925 | |
| ShapeLLMParameters=7B2026.05 | 30.075 | 11.491 | 1.222 | 14.035 | 13.316 | 37.693 | 38.047 | 24.56 | |
| ShapeLLM-Omni2026.05 | 20.19 | 11.326 | 1.197 | 14.19 | 13.276 | 34.617 | 35.115 | 25.625 | |
| MiniGPT-3D2026.05 | 7.56 | 5.482 | 0.468 | 7.68 | 10.531 | 28.708 | 23.722 | 17.36 | |
| LLaMA-Mesh2026.05 | 7.085 | 5.119 | 0.441 | 7.439 | 9.904 | 26.326 | 25.095 | 10.74 |