Talking Face Generation on HDTF
0.0103PDFreeTalkDiff
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| FreeTalkDiffSetting=Few-shot, Model Architecture=Diffusion-based2026.05 | 0.0103 | 0.883 | 0.718 | 0.5 | 0.023 | 0.289 | |
| LatentSyncSetting=Few-shot, Model Architecture=Diffusion-based2026.05 | 0.0148 | 0.827 | 0.54 | 1.5 | 0.025 | 0.255 | |
| Hallo2Setting=One-shot, Model Architecture=Diffusion-based2026.05 | 0.0152 | 0.844 | 0.547 | 3.9 | 0.191 | 0.211 | |
| EchoMimicSetting=One-shot, Model Architecture=Diffusion-based2026.05 | 0.016 | 0.84 | 0.524 | 50.1 | 0.476 | 0.145 | |
| SonicSetting=One-shot, Model Architecture=Diffusion-based2026.05 | 0.0164 | 0.832 | 0.558 | 23.8 | 0.415 | 0.153 | |
| SadTalkerSetting=One-shot, Model Architecture=Non-diffusion-based2026.05 | 0.0165 | 0.816 | 0.544 | 35.2 | 0.225 | 0.212 | |
| TalkLipSetting=Few-shot, Model Architecture=Non-diffusion-based2026.05 | 0.0174 | 0.805 | 0.424 | 14 | 0.06 | 0.201 | |
| MuseTalkSetting=Few-shot, Model Architecture=Diffusion-based2026.05 | 0.0179 | 0.788 | 0.343 | 2.4 | 0.037 | 0.241 | |
| Audio2HeadSetting=One-shot, Model Architecture=Non-diffusion-based2026.05 | 0.0179 | 0.76 | 0.282 | 92.2 | 0.314 | 0.058 | |
| Wav2LipSetting=Few-shot, Model Architecture=Non-diffusion-based2026.05 | 0.019 | 0.756 | 0.358 | 12.4 | 0.057 | 0.205 | |
| MakeItTalkSetting=One-shot, Model Architecture=Non-diffusion-based2026.05 | 0.0244 | 0.752 | 0.25 | 87.5 | 0.293 | 0.042 |