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/

Zendy Plus

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

Zendy Tools

Zendy Open

Diffusion models have become a mainstream approach for high-resolution imagesynthesis. However, directly generating higher-resolution images frompretrained diffusion models will encounter unreasonable object duplication andexponentially increase the generation time. In this paper, we discover thatobject duplication arises from feature duplication in the deep blocks of theU-Net. Concurrently, We pinpoint the extended generation times toself-attention redundancy in U-Net's top blocks. To address these issues, wepropose a tuning-free higher-resolution framework named HiDiffusion.Specifically, HiDiffusion contains Resolution-Aware U-Net (RAU-Net) thatdynamically adjusts the feature map size to resolve object duplication andengages Modified Shifted Window Multi-head Self-Attention (MSW-MSA) thatutilizes optimized window attention to reduce computations. we can integrateHiDiffusion into various pretrained diffusion models to scale image generationresolutions even to 4096x4096 at 1.5-6x the inference speed of previousmethods. Extensive experiments demonstrate that our approach can address objectduplication and heavy computation issues, achieving state-of-the-artperformance on higher-resolution image synthesis tasks.

HiDiffusion: Unlocking Higher-Resolution Creativity and Efficiency in Pretrained Diffusion Models