작성자: 16기 천원준
https://arxiv.org/abs/2006.11239
Denoising Diffusion Probabilistic Models
We present high quality image synthesis results using diffusion probabilistic models, a class of latent variable models inspired by considerations from nonequilibrium thermodynamics. Our best results are obtained by training on a weighted variational bound
arxiv.org
Forward/Diffusion process
Reverse process
Training
Forward process and LT
Reverse process and L1:T-1
Data scaling, reverse process decoder and L0
Simplified Training objective
댓글 영역