Visual Planning on VBench RealEstate10K (val)
93.7Subject ConsistencyAether
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| AetherLatency(s)=191.41, FLOPs(T)=14352.50, Speed=1.00×2025.08 | 93.7 | 95.21 | 98.03 | 90.09 | 49.33 | 64.21 | |
| HEROBase Model=Aether, M=2, Latency(s)=113.98, FLOPs(T)=8295.21, Speed=1.73×2025.08 | 93.42 | 95.02 | 97.9 | 90.09 | 49.34 | 64.21 | |
| CogVideoX-I2V2025.08 | 92.72 | 93.87 | 98.94 | 61.26 | 50.79 | 68.64 | |
| HEROBase Model=Aether, M=3, Latency(s)=101.13, FLOPs(T)=7241.78, Speed=1.97×2025.08 | 92.05 | 94.9 | 97.53 | 90.09 | 50.02 | 61.92 | |
| ToCaBase Model=Aether, M=2, Latency(s)=138.45, FLOPs(T)=10053.38, Speed=1.43×2025.08 | 92.03 | 94.37 | 97.86 | 90.09 | 49.68 | 60.76 | |
| FoRABase Model=Aether, M=2, Latency(s)=108.43, FLOPs(T)=7753.18, Speed=1.85×2025.08 | 91.94 | 94.53 | 97.58 | 90.09 | 47.62 | 61.9 | |
| TaylorSeerBase Model=Aether, M=2, Latency(s)=108.43, FLOPs(T)=7753.18, Speed=1.85×2025.08 | 89.06 | 93.62 | 96.23 | 89.18 | 47.78 | 58.55 |