2 ms·
Temporally Coherent Human Video Deepfakes via Diffusion
- fritzschopen 4y agoWhere's the code? Looks a bit suspect imo. Generating human motion without being driven by a real-world video is interesting tho. But adding audio speech as a driver. Well...