search

fudan-generative-vision / hallo

Hallo: Hierarchical Audio-Driven Visual Synthesis for Portrait Image Animation
https://fudan-generative-vision.github.io/hallo/
MIT License
8.24k stars 1.1k forks source link

Facial features get distort #138

Open rishabh-cruv opened 1 month ago

rishabh-cruv commented 1 month ago

After running inference for audio file of 47 sec with image in which speaker had wore glasses and had plain background. Starting was fine but at the end, facial features got distorted (beard growth, skin tone/color changes, background color changes)

rishabh-cruv commented 1 month ago

Also, how can I decrease blink delay?