Dynamics Prediction on Multiple-objects Real-world
0.0058MSEDiffusion-UNet w/ Contact
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Diffusion-UNet w/ ContactTraining Regime=Real-Finetune2026.01 | 0.0058 | 79.12 | |
| Diffusion-UNetTraining Regime=Real-Finetune2026.01 | 0.0065 | 75.82 | |
| MLPTraining Regime=Real-Finetune2026.01 | 0.0069 | 73.43 | |
| Diffusion-UNet w/ ContactTraining Regime=Real data2026.01 | 0.0075 | 73.33 | |
| MLPTraining Regime=Real data2026.01 | 0.0082 | 71.76 | |
| Diffusion-UNetTraining Regime=Real data2026.01 | 0.0083 | 73.04 | |
| UNetTraining Regime=Real data2026.01 | 0.0084 | 72.09 | |
| UNetTraining Regime=Real-Finetune2026.01 | 0.0085 | 74.12 | |
| BaselineTraining Regime=Real data2026.01 | 0.0161 | 62.98 |