返回文献探索AI 摘要
Paper · arXiv 2402.13929
SDXL-Lightning: Progressive Adversarial Diffusion Distillation
Shanchuan Lin, Anran Wang, Xiao Yang
27 upvotesFebruary 21, 2024arXiv 预印本
A diffusion distillation method using progressive and adversarial techniques achieves high-quality one-step/few-step text-to-image generation with improved mode coverage and is available as a distilled model.
diffusion distillationprogressive distillationadversarial distillationSDXLtext-to-image generationdiscriminator designLoRAUNetmode coverage
Abstract
We propose a diffusion distillation method that achieves new state-of-the-art in one-step/few-step 1024px text-to-image generation based on SDXL. Our method combines progressive and adversarial distillation to achieve a balance between quality and mode coverage. In this paper, we discuss the theoretical analysis, discriminator design, model formulation, and training techniques. We open-source our distilled SDXL-Lightning models both as LoRA and full UNet weights.