On Equivariance and Fast Sampling in Video Diffusion Models Trained with Warped Noise

Open in new window