Audio-driven facial animation on 100 examples (test)
200Time To First BoundaryHan et al.
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Han et al.2026.05 | 200 | 200 | 1.4 | 401.4 | |
| Lee et al.2026.05 | 200 | 200 | 10 | 410 | |
| DiffPoseTalkHardware=Nvidia L40S GPU2026.05 | 200 | 200 | 3,326 | 3,726 | |
| TokTalkHardware=Nvidia L40S GPU, Flow matching steps (n)=1 to 202026.05 | 200 | 200 | 26 | 400 |