Text-to-3D Generation on Toys4K CL (Base)
29.6CLIP SimilarityBase Training
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Base TrainingBackbone=TRELLIS-XL (Flow-based)2026.04 | 29.6 | 73.04 | 75.22 | |
| Joint TrainingBackbone=TRELLIS-XL (Flow-based)2026.04 | 29.57 | 78.56 | 74.34 | |
| Base TrainingBackbone=Shap-E (Diffusion-based)2026.04 | 28.75 | 107.11 | 25.39 | |
| Joint TrainingBackbone=Shap-E (Diffusion-based)2026.04 | 28.45 | 112.23 | 25.65 | |
| ReConText3D + L2-SPBackbone=TRELLIS-XL (Flow-based)2026.04 | 28.44 | 84.85 | 71.57 | |
| ReConText3DBackbone=TRELLIS-XL (Flow-based)2026.04 | 28.41 | 87.28 | 70.2 | |
| ReConText3DBackbone=Shap-E (Diffusion-based)2026.04 | 28.34 | 110.49 | 25.21 | |
| ReConText3D + L2-SPBackbone=Shap-E (Diffusion-based)2026.04 | 28.33 | 112.31 | 25.25 | |
| Fine-tuningBackbone=Shap-E (Diffusion-based)2026.04 | 27.8 | 118.28 | 25.46 | |
| L2-SPBackbone=Shap-E (Diffusion-based)2026.04 | 27.38 | 123.98 | 25.76 | |
| L2-SPBackbone=TRELLIS-XL (Flow-based)2026.04 | 24.5 | 102.88 | 68.47 | |
| Fine-tuningBackbone=TRELLIS-XL (Flow-based)2026.04 | 24.46 | 101 | 72.01 |