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

Stable Diffusion XL (SDXL) has become the best open source text-to-imagemodel (T2I) for its versatility and top-notch image quality. Efficientlyaddressing the computational demands of SDXL models is crucial for wider reachand applicability. In this work, we introduce two scaled-down variants, SegmindStable Diffusion (SSD-1B) and Segmind-Vega, with 1.3B and 0.74B parameterUNets, respectively, achieved through progressive removal using layer-levellosses focusing on reducing the model size while preserving generative quality.We release these models weights at https://hf.co/Segmind. Our methodologyinvolves the elimination of residual networks and transformer blocks from theU-Net structure of SDXL, resulting in significant reductions in parameters, andlatency. Our compact models effectively emulate the original SDXL bycapitalizing on transferred knowledge, achieving competitive results againstlarger multi-billion parameter SDXL. Our work underscores the efficacy ofknowledge distillation coupled with layer-level losses in reducing model sizewhile preserving the high-quality generative capabilities of SDXL, thusfacilitating more accessible deployment in resource-constrained environments.

Progressive Knowledge Distillation Of Stable Diffusion XL Using Layer Level Loss