Video Generation on UCF-101 64x64 (test)
158.7FVDSCD-B^D
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SCD-B^DSec/F=2.8, config_variant=Increased decoder depth2026.02 | 158.7 | 0.036 | 0.829 | 26 | |
| SCD-B^ESec/F=1.1, config_variant=Increased encoder depth2026.02 | 171.1 | 0.037 | 0.829 | 25.98 | |
| SCD-BSec/F=1.12026.02 | 174.7 | 0.038 | 0.824 | 25.78 | |
| Causal DiT-BSec/F=3.92026.02 | 187.6 | 0.038 | 0.827 | 25.85 | |
| FAR-BSec/F=3.22026.02 | 194.1 | 0.037 | 0.818 | 25.64 | |
| RaMViDSec/F=non-causal2026.02 | 396.7 | 0.09 | 0.639 | 21.37 | |
| MCVD-cpSec/F=non-causal2026.02 | 468.1 | 0.088 | 0.658 | 21.82 | |
| DIGANtrained on k=16, Resolution=64 x 642022.05 | 655 | — | — | — | |
| MCVD spatin past-masktrained on k=4, Resolution=64 x 642022.05 | 1,143 | — | — | — | |
| TGANv2trained on k=16, Resolution=64 x 642022.05 | 1,209 | — | — | — | |
| MCVD concat past-masktrained on k=4, Resolution=64 x 642022.05 | 1,228.3 | — | — | — | |
| MoCoGAN-MDPtrained on k=16, Resolution=64 x 642022.05 | 1,277 | — | — | — |