SD3.5-Flash：基于分布引导的生成流蒸馏技术

摘要

我们推出SD3.5-Flash，一种高效的少步蒸馏框架，旨在将高质量图像生成能力引入普及型消费设备。该方法通过专门为少步生成重新设计的分布匹配目标，对计算量巨大的校正流模型进行蒸馏。我们引入两项关键创新：“时间步共享”以减少梯度噪声，以及“分步微调”以提升提示对齐效果。结合文本编码器重构和专用量化等全面的管道优化措施，我们的系统实现了快速生成和跨不同硬件配置的内存高效部署。这使从手机到台式机的全系列设备都能平等地获得这一技术。通过包括大规模用户研究在内的广泛评估，我们证明SD3.5-Flash在少步方法中持续领先，使先进的生成式AI真正适用于实际部署。

English

We present SD3.5-Flash, an efficient few-step distillation framework that brings high-quality image generation to accessible consumer devices. Our approach distills computationally prohibitive rectified flow models through a reformulated distribution matching objective tailored specifically for few-step generation. We introduce two key innovations: "timestep sharing" to reduce gradient noise and "split-timestep fine-tuning" to improve prompt alignment. Combined with comprehensive pipeline optimizations like text encoder restructuring and specialized quantization, our system enables both rapid generation and memory-efficient deployment across different hardware configurations. This democratizes access across the full spectrum of devices, from mobile phones to desktop computers. Through extensive evaluation including large-scale user studies, we demonstrate that SD3.5-Flash consistently outperforms existing few-step methods, making advanced generative AI truly accessible for practical deployment.

SD3.5-Flash：基于分布引导的生成流蒸馏技术

SD3.5-Flash: Distribution-Guided Distillation of Generative Flows

摘要

Support