Talking Face Generation on HDTF (long)
118.97FVDLip Forcing
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Lip ForcingModel parameter size=14B2026.06 | 118.97 | 4.27 | 0.9216 | 0.945 | 6.86 | 8.68 | |
| Lip ForcingModel parameter size=1.3B2026.06 | 133.41 | 23.98 | 0.8797 | 0.9035 | 7.87 | 7.53 | |
| MuseTalk2026.06 | 143.2 | 6.89 | 0.939 | 0.9583 | 6.87 | 8.5 | |
| LatentSync2026.06 | 150.51 | 4.04 | 0.9399 | — | 5.92 | 9.98 | |
| Diff2Lip2026.06 | 311.98 | 18.32 | 0.7285 | 0.8749 | 6.61 | 8.67 | |
| VideoReTalking2026.06 | 313.4 | 21.53 | 0.8711 | 0.9057 | 6.72 | 8.7 | |
| Wav2Lip2026.06 | 372.47 | 20.26 | 0.9045 | 0.9434 | 6.56 | 9.17 |