Long-horizon Video Generation on RealEstate10K Revisiting
21.85PSNROurs-Wan
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Ours-WanBase Model=Wan2026.04 | 21.85 | 71 | 0.2 | 51.24 | 72.27 | 98.86 | 97.62 | |
| Ours-CogBase Model=CogVideoX2026.04 | 20.23 | 69 | 0.23 | 50.64 | 63.53 | 99.17 | 92.86 | |
| WorldPlayVariant=bidirectional attention2026.04 | 16.31 | 64 | 0.36 | 51.51 | 70.45 | 99.06 | 92.86 | |
| VMem2026.04 | 15.37 | 65 | 0.37 | 42.34 | 59.8 | 97.66 | 78.57 | |
| WorldPlay-dVariant=distilled2026.04 | 15.09 | 63 | 0.31 | 51.34 | 71.43 | 99.32 | 88.1 | |
| DFoT2026.04 | 14.93 | 48 | 0.34 | 43 | 66.8 | 98.58 | 35.71 | |
| SEVA2026.04 | 14.8 | 60 | 0.47 | 42.78 | 61.67 | 98.4 | 88.1 | |
| AC3DBase Model=CogVideoX-5B-I2V2026.04 | 13.13 | 52 | 0.52 | 51.37 | 64.71 | 99.29 | 95.24 |