FireFlow：用于图像语义编辑的快速矫正流反演

FireFlow: Fast Inversion of Rectified Flow for Image Semantic Editing

December 10, 2024

作者: Yingying Deng, Xiangyu He, Changwang Mei, Peisong Wang, Fan Tang

cs.AI

摘要

尽管具有蒸馏的矫正流（ReFlows）提供了一种快速取样的有前途的方法，但其快速反演将图像转换回结构化噪音以进行恢复和随后的编辑问题尚未解决。本文介绍了FireFlow，这是一种简单而有效的零样本方法，它继承了基于ReFlow的模型（如FLUX）在生成方面的惊人能力，同时将其能力扩展到准确的反演和编辑中，需要8个步骤。我们首先证明，一个精心设计的数值求解器对于ReFlow反演至关重要，可以实现准确的反演和重建，具有二阶求解器的精度，同时保持一阶欧拉方法的实际效率。与最先进的ReFlow反演和编辑技术相比，该求解器实现了3倍的运行时加速，同时在无需训练的模式下提供更小的重建误差和更优秀的编辑结果。代码可在https://github.com/HolmesShuan/FireFlow{此URL}找到。

English

Though Rectified Flows (ReFlows) with distillation offers a promising way for fast sampling, its fast inversion transforms images back to structured noise for recovery and following editing remains unsolved. This paper introduces FireFlow, a simple yet effective zero-shot approach that inherits the startling capacity of ReFlow-based models (such as FLUX) in generation while extending its capabilities to accurate inversion and editing in 8 steps. We first demonstrate that a carefully designed numerical solver is pivotal for ReFlow inversion, enabling accurate inversion and reconstruction with the precision of a second-order solver while maintaining the practical efficiency of a first-order Euler method. This solver achieves a 3times runtime speedup compared to state-of-the-art ReFlow inversion and editing techniques, while delivering smaller reconstruction errors and superior editing results in a training-free mode. The code is available at https://github.com/HolmesShuan/FireFlow{this URL}.