Talking Head Generation on CelebV-HQ (test)
6.818FIDArchon
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Archontrained on the benchmark dataset=false2026.05 | 6.818 | 93.81 | 5.21 | 8.998 | — | — | — | 3.794 | |
| Hallo3trained on the benchmark dataset=false2026.05 | 15.67 | 105.5 | 5.429 | 9.158 | — | — | — | 3.722 | |
| AniPortraittrained on the benchmark dataset=true2026.05 | 39.73 | 160.7 | 3.493 | 10.982 | — | — | — | 3.833 | |
| SonicMethod Type=Diffusion-based, Runtime=17.042024.11 | 43.137 | 483.108 | 2.689 | 10.194 | 1.783 | 0.9624 | 0.9972 | — | |
| HalloMethod Type=Diffusion-based, Runtime=74.652024.11 | 47.403 | 488.499 | 2.68 | 10.292 | 2.273 | 0.9607 | 0.9942 | — | |
| EchoMimicMethod Type=Diffusion-based, Runtime=5.452024.11 | 48.267 | 596.87 | 1.949 | 10.754 | 2.136 | 0.9612 | 0.9932 | — | |
| Hallo2Method Type=Diffusion-based, Runtime=45.752024.11 | 52.396 | 481.336 | 2.638 | 10.343 | 2.819 | 0.9587 | 0.9942 | — | |
| AniPortraitMethod Type=Diffusion-based, Runtime=44.032024.11 | 53.746 | 590.373 | 0.996 | 12.084 | 3.296 | 0.9522 | 0.9911 | — | |
| EchoMimictrained on the benchmark dataset=true2026.05 | 56.81 | 236.9 | 4.463 | 9.575 | — | — | — | 3.601 | |
| SadTalkerMethod Type=GAN-based, Runtime=3.752024.11 | 57.574 | 841.962 | 1.978 | 10.915 | 2.252 | 0.9434 | 0.9959 | — | |
| V-ExpressMethod Type=Diffusion-based, Runtime=39.042024.11 | 65.4 | 889.985 | 0.809 | 13.255 | 2.713 | 0.9019 | 0.9965 | — |