Improved Denoising Diffusion Probabilistic Models
Alex Nichol * 1 Prafulla Dhariwal * 1
Abstract and VAEs (Kingma & Welling, 2013). This raises various
Denoising diffusion probabilistic models (DDPM) questions, such as whether DDPMs are capable of capturing
are a class of generative models which have re- all the modes of a distribution. Furthermore, while Ho et al.
cently been shown to produce excellent sam- (2020) ...


雷达卡




京公网安备 11010802022788号







