Talking Face Generation on CREMA
0.0086PDFreeTalkDiff
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| FreeTalkDiffSetting=Few-shot, Model Architecture=Diffusion-based2026.05 | 0.0086 | 0.887 | 0.711 | 1.5 | 0.02 | 0.218 | |
| TalkLipSetting=Few-shot, Model Architecture=Non-diffusion-based2026.05 | 0.014 | 0.783 | 0.475 | 15.4 | 0.065 | 0.146 | |
| Audio2HeadSetting=One-shot, Model Architecture=Non-diffusion-based2026.05 | 0.0151 | 0.807 | 0.531 | 80.2 | 0.313 | 0.058 | |
| LatentSyncSetting=Few-shot, Model Architecture=Diffusion-based2026.05 | 0.0177 | 0.762 | 0.429 | 2.2 | 0.032 | 0.189 | |
| SonicSetting=One-shot, Model Architecture=Diffusion-based2026.05 | 0.0177 | 0.707 | 0.408 | 32.9 | 0.228 | 0.104 | |
| MuseTalkSetting=Few-shot, Model Architecture=Diffusion-based2026.05 | 0.0185 | 0.668 | 0.313 | 5.2 | 0.037 | 0.143 | |
| EchoMimicSetting=One-shot, Model Architecture=Diffusion-based2026.05 | 0.0193 | 0.665 | 0.31 | 71 | 0.291 | 0.081 | |
| Wav2LipSetting=Few-shot, Model Architecture=Non-diffusion-based2026.05 | 0.0195 | 0.649 | 0.235 | 14.9 | 0.056 | 0.134 | |
| MakeItTalkSetting=One-shot, Model Architecture=Non-diffusion-based2026.05 | 0.0202 | 0.745 | 0.417 | 74.1 | 0.213 | 0.023 | |
| Hallo2Setting=One-shot, Model Architecture=Diffusion-based2026.05 | 0.0203 | 0.627 | 0.121 | 7.9 | 0.175 | 0.131 | |
| SadTalkerSetting=One-shot, Model Architecture=Non-diffusion-based2026.05 | 0.0244 | 0.666 | 0.337 | 61.6 | 0.294 | 0.108 |