Text-to-3D Generation on Text-to-3D Generation Prompts Quantitative Study
31.1CLIP ScoreBaseline
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| BaselineDistillation Objective=ISM, Framework=LucidDreamer2026.05 | 31.1 | 0.413 | 0.225 | 6.438 | 1,706 | 110.6 | |
| BaselineDistillation Objective=SDS, Framework=GaussianDreamer2026.05 | 30.14 | -0.125 | 0.194 | 9.908 | 1,067 | 46.4 | |
| CAdamDistillation Objective=SDS, Framework=GaussianDreamer2026.05 | 29.95 | -0.028 | 0.185 | 9.643 | 143 | 6.1 | |
| BaselineDistillation Objective=VFDS, Framework=FlowDreamer2026.05 | 29.88 | 0.436 | 0.235 | 7.183 | 3,373 | 218.8 | |
| CAdamDistillation Objective=VFDS, Framework=FlowDreamer2026.05 | 29.21 | 0.486 | 0.237 | 8.891 | 82 | 5.3 | |
| CAdamDistillation Objective=ISM, Framework=LucidDreamer2026.05 | 28.48 | 0.518 | 0.225 | 8.167 | 81 | 5.3 |