Diffusion-RWKV: Scaling RWKV-Like Architectures for Diffusion Models

arXiv:2404.04478v1 Announce Type: new
Abstract: Transformers have catalyzed advancements in computer vision and natural language processing (NLP) fields. However, substantial computational complexity poses limitations for their application in long-context tasks, such as high-resolution image generation. This paper introduces a series of architectures adapted from the RWKV model used in the NLP, with requisite modifications tailored for diffusion model applied to image generation tasks, referred to as Diffusion-RWKV. Similar to the diffusion with Transformers, our model is designed to efficiently handle …

abstract and natural language processing application architectures arxiv complexity computational computer computer vision context cs.cv diffusion diffusion models fields however image image generation language language processing limitations natural natural language natural language processing nlp paper processing resolution rwkv scaling series tasks transformers type vision