A new one shot face swap approach for image and video domains
Generator model with unet backbone and Discriminator model for 1, 2, 3 blocks in AAD ResBlock
InsightFace arcface model