English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Text-to-image generation has recently witnessed remarkable achievements. Weintroduce a text-conditional image diffusion model, termed RAPHAEL, to generatehighly artistic images, which accurately portray the text prompts, encompassingmultiple nouns, adjectives, and verbs. This is achieved by stacking tens ofmixture-of-experts (MoEs) layers, i.e., space-MoE and time-MoE layers, enablingbillions of diffusion paths (routes) from the network input to the output. Eachpath intuitively functions as a "painter" for depicting a particular textualconcept onto a specified image region at a diffusion timestep. Comprehensiveexperiments reveal that RAPHAEL outperforms recent cutting-edge models, such asStable Diffusion, ERNIE-ViLG 2.0, DeepFloyd, and DALL-E 2, in terms of bothimage quality and aesthetic appeal. Firstly, RAPHAEL exhibits superiorperformance in switching images across diverse styles, such as Japanese comics,realism, cyberpunk, and ink illustration. Secondly, a single model with threebillion parameters, trained on 1,000 A100 GPUs for two months, achieves astate-of-the-art zero-shot FID score of 6.61 on the COCO dataset. Furthermore,RAPHAEL significantly surpasses its counterparts in human evaluation on theViLG-300 benchmark. We believe that RAPHAEL holds the potential to propel thefrontiers of image generation research in both academia and industry, pavingthe way for future breakthroughs in this rapidly evolving field. More detailscan be found on a webpage: https://raphael-painter.github.io/.

RAPHAEL: Text-to-Image Generation via Large Mixture of Diffusion Paths