[Lightspeed]

Deepfake SotA:

Few-Shot Adversarial Learning of Realistic Neural Talking Head Models