Year: 2023
Type: preprint
Abstract: Recent ODE/SDE-based generative models, such as diffusion models, rectified flows, and flow matching, define a generative process as a time reversal of a fixed forward process. Even though these models show impressive performance on large-scale datasets, numerical simulation requires multiple evaluations of a neural network, leading to a slow sampling speed. We attribute the reason to the high curvature of the learned generative trajectories, as it is directly related to the truncation error of a numerical solver. Based on the relationship between the forward process and the curvature, here we present an efficient method of training the forward process to minimize the curvature of generative trajectories without any ODE/SDE simulation. Experiments show that our method achieves a lower curvature than previous models and, therefore, decreased sampling costs while maintaining competitive performance. Code is available at https://github.com/sangyun884/fast-ode.
Source: arXiv (Cornell University)
Authors: Lee, Sangyun, Kim, Beomsu, Ye, Jong Chul
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funders: Korea Medical Device Development Fund, National Research Foundation, Ministry of Trade, Industry and Energy, Ministry of Food and Drug Safety, National Research Foundation of Korea, Korea Advanced Institute of Science and Technology, Iran Telecommunication Research Center, Ministry of Science and ICT, South Korea, Korea Customs Service, Institute for Information and Communications Technology Promotion
OpenAlex ID: W4318719691