Text-to-Video Generation on MSR-VTT (val)
0.301Frame Consistency (F.C.)VC2+CV-VAE-V
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VC2+CV-VAE-VFCR=4x, Trainable=true, Resolution=320x5122024.05 | 0.301 | 0.987 | |
| VC2FCR=1x, Trainable=false, Resolution=320x5122024.05 | 0.292 | 0.96 | |
| VC2+CV-VAE-IFCR=1x, Trainable=false, Resolution=320x5122024.05 | 0.29 | 0.964 | |
| VC2+CV-VAE-VFCR=4x, Trainable=false, Resolution=320x5122024.05 | 0.285 | 0.953 |