Diffusion models have shown great promise for image and video generation, but sampling from state-of-the-art models requires expensive numerical integration of a generative ODE. One approach for tackling this problem is rectified flows, which iteratively learn smooth ODE paths that are less susceptible to truncation error. However, rectified flows still require a relatively large number of function evaluations (NFEs). In this work, we propose improved techniques for training rectified flows, allowing them to compete with \emph{knowledge distillation} methods even in the low NFE setting. Our main insight is that under realistic settings, a single iteration of the Reflow algorithm for training rectified flows is sufficient to learn nearly straight trajectories; hence, the current practice of using multiple Reflow iterations is unnecessary. We thus propose techniques to improve one-round training of rectified flows, including a U-shaped timestep distribution and LPIPS-Huber premetric. With these techniques, we improve the FID of the previous 2-rectified flow by up to 75\% in the 1 NFE setting on CIFAR-10. On ImageNet 64$\times$64, our improved rectified flow outperforms the state-of-the-art distillation methods such as consistency distillation and progressive distillation in both one-step and two-step settings and rivals the performance of improved consistency training (iCT) in FID. Code is available at https://github.com/sangyun884/rfpp.
Delta Rectified Flow Sampling for Text-to-Image Editing
Harvard AI and Robotics Lab, Harvard University(哈佛大学人工智能与机器人实验室);Computer Science Department, Harvard University(哈佛大学计算机科学系);Multimodal Intelligence and Perception Lab, DGIST(大邱庆北科学技术院多模态智能与感知实验室);Kempner Institute for the Study of Natural and Artificial Intelligence, Harvard University(哈佛大学肯普纳自然与人工智能研究所)
StreamFlow: Theory, Algorithm, and Implementation for High-Efficiency Rectified Flow Generation
Rutgers University, New Jersey, USA(罗杰斯大学);Georgia Institute of Technology, Atlanta, Georgia, USA(佐治亚理工学院);Nanyang Technological University, Singapore(南洋理工大学);University of Wisconsin-Madison, Wisconsin, USA(威斯康星大学麦迪逊分校)
Erased, But Not Forgotten: Erased Rectified Flow Transformers Still Remain Unsafe Under Concept Attack
Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing, School of Artificial Intelligence, Beihang University(北京航空航天大学人工智能学院未来区块链与隐私计算北京高精尖创新中心);University of Science and Technology of China(中国科学技术大学);National University of Defense Technology(国防科技大学);GigaAI(极佳科技);National University of Singapore(新加坡国立大学)
I-Max: Maximize the Resolution Potential of Pre-trained Rectified Flow Transformers with Projected Flow
Beijing University of Posts and Telecommunications(北京邮电大学);Shanghai Artificial Intelligence Laboratory(上海人工智能实验室);The Chinese University of Hong Kong(香港中文大学)
SLayR: Scene Layout Generation with Rectified Flow
Saarland University(萨尔大学);Max Planck Institute for Informatics(马克斯·普朗克信息学研究所);Max Planck Institute for Software Systems(马克斯·普朗克软件系统研究所);University of Technology Nuremberg(纽伦堡工业大学)
Offline Preference Optimization for Rectified Flow with Noise-Tracked Pairs
Zhejiang University(浙江大学);Shanghai Institute for Advanced Study-Zhejiang University(上海先进研究院-浙江大学);Shanghai Institute for Mathematics and Interdisciplinary Sciences(上海数学与交叉科学研究院)
TReFT: Taming Rectified Flow Models For One-Step Image Translation
University of Chinese Academy of Sciences(中国科学院大学);Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所);Beihang University(北航);Tianjin University(天津大学);CreateAI