SadTalker v0.0.2 Release Note

In v0.0.2, we finally solved the problems of paste back and got a more stable face with the following options using a newly trained model!

python inference.py --still --enhancer gfpgan --preprocess full

full3.eluosi_enhanced.mp4

full_body_2.bus_chinese_enhanced.mp4

full_body_1.bus_chinese_enhanced.mp4

Have fun and thanks for all your support! ❤️