Redirecting to original paper in 30 seconds...
Click below to go immediately or wait for automatic redirect
📄 Abstract
Abstract: Rectified Flow offers a simple and effective approach to high-quality
generative modeling by learning a velocity field. However, we identify a
limitation in directly modeling the velocity with an unconstrained neural
network: the learned velocity often fails to satisfy certain boundary
conditions, leading to inaccurate velocity field estimations that deviate from
the desired ODE. This issue is particularly critical during stochastic sampling
at inference, as the score function's errors are amplified near the boundary.
To mitigate this, we propose a Boundary-enforced Rectified Flow Model (Boundary
RF Model), in which we enforce boundary conditions with a minimal code
modification. Boundary RF Model improves performance over vanilla RF model,
demonstrating 8.01% improvement in FID score on ImageNet using ODE sampling and
8.98% improvement using SDE sampling.
Authors (8)
Xixi Hu
Runlong Liao
Keyang Xu
Bo Liu
Yeqing Li
Eugene Ie
+2 more
Key Contributions
Proposes a Boundary-enforced Rectified Flow Model (Boundary RF Model) that enforces boundary conditions on the learned velocity field. This addresses limitations in vanilla Rectified Flow where unconstrained networks lead to inaccurate velocity fields, especially near boundaries, improving sample quality.
Business Value
Enables the generation of higher-fidelity synthetic data (e.g., images) for training other models, creative applications, or data augmentation, leading to better performance and reduced data acquisition costs.