Audio-Visual Generation on VGGSound SS
615.28FVDBridgeDiT
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| BridgeDiTProcess=T → VA2025.10 | 615.28 | 6.01 | 36.2 | |
| JavisDiTProcess=T → VA2025.10 | 637.5 | 7.78 | 17.9 | |
| Wan + SDAProcess=T → V ∥ T → A2025.10 | 737.96 | 8.05 | 23.8 | |
| Wan + MMAudioProcess=T → V + V → A2025.10 | 737.96 | 5.39 | 33.3 | |
| Wan + SeeingHearingProcess=T → V + V → A2025.10 | 737.96 | 8.02 | 32.1 | |
| JointDiTProcess=T → I + I → VA2025.10 | 866.59 | 7.19 | 12.5 | |
| SSVGProcess=T → VA2025.10 | 1,032.87 | 8.86 | 14.8 | |
| CoDiProcess=T → VA2025.10 | 1,203.68 | 8.38 | 11.3 | |
| SDA + TPosProcess=T → A + A → V2025.10 | 1,732.47 | 8.05 | 16.3 | |
| SDA + TempoTokenProcess=T → A + A → V2025.10 | 1,942.94 | 8.05 | 24.2 |