R2-Talker adopts the motion generator of GeneFace++ [45] which generates 3D landmarks from input driving audio？

KylinYee / R2-Talker-code

R2-Talker: Realistic Real-Time Talking Head Synthesis with Hash Grid Landmarks Encoding and Progressive Multilayer Conditioning

MIT License

71 stars 4 forks source link

Open kike-0304 opened 3 months ago

kike-0304 commented 3 months ago

how to adopts the motion generator of GeneFace++ [45] which generates 3D landmarks from input driving audio？