Everything at Every Scale: Scale-Invariant Diffusion with Continuous Super-Resolution
Zixin Jessie Chen, Zhuo Chen, Archer Wang, Jeff Gore, William T. Freeman, Congyue Deng, Marin Soljačić
Read on arXiv →Key claim
SKILD unifies image generation and super-resolution effectively.
The SKILD model introduces a unified framework for image generation and super-resolution, achieving impressive results on CIFAR-10 and ImageNet. It operates without task-specific architectures or retraining, making it a versatile tool for image processing.
In plain English
The SKILD model introduces a unified framework for image generation and super-resolution, achieving impressive results on CIFAR-10 and ImageNet. It operates without task-specific architectures or retraining, making it a versatile tool for image processing.
The SKILD model unifies image generation and super-resolution in a novel way, leveraging scale invariance.
The empirical results on CIFAR-10 and ImageNet are strong, with clear performance metrics provided.