Self-Rectifying Diffusion Sampling with Perturbed-Attention Guidance

arXiv:2403.17377v1 Announce Type: cross
Abstract: Recent studies have demonstrated that diffusion models are capable of generating high-quality samples, but their quality heavily depends on sampling guidance techniques, such as classifier guidance (CG) and classifier-free guidance (CFG). These techniques are often not applicable in unconditional generation or in various downstream tasks such as image restoration. In this paper, we propose a novel sampling guidance, called Perturbed-Attention Guidance (PAG), which improves diffusion sample quality across both unconditional and conditional settings, achieving this …

abstract arxiv attention classifier cs.ai cs.cv cs.lg diffusion diffusion models free guidance image image restoration quality samples sampling studies tasks type