SDXL-Lightning：渐进式对抗扩散蒸馏

SDXL-Lightning: Progressive Adversarial Diffusion Distillation

February 21, 2024

作者: Shanchuan Lin, Anran Wang, Xiao Yang

cs.AI

摘要

我们提出了一种扩散蒸馏方法，基于SDXL，在一步/少步1024像素文本到图像生成中取得了新的最先进成果。我们的方法结合了渐进式和对抗性蒸馏，以在质量和模式覆盖之间取得平衡。本文讨论了理论分析、鉴别器设计、模型公式和训练技术。我们以LoRA和完整UNet权重的形式开源了我们蒸馏的SDXL-Lightning模型。

English

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.