FLOAT: Generative Motion Latent Flow Matching for Audio-driven Talking Portrait Paper • 2412.01064 • Published Dec 2, 2024 • 47
