Duplicated from hysts/ControlNet
I wonder how you use ControlNet for video generation? Just preprocess original video and send the sequence of preprocessed frames as condition?
The input noises of Stable Diffusion are initialized the same for all frames?
· Sign up or log in to comment