Stochastic Sampling from Deterministic Flow Models
Keywords: rectified flow, diffusion models, stochastic sampling, generative model
TL;DR: We enable stochastic sampling of popular deterministic generative models like rectified flows and probability flow ODEs
Abstract: Deterministic flow models such as rectified flows offer a general framework for learning a deterministic transport map between two distributions, realized as the vector field for an ordinary differential equation (ODE). However, they are sensitive to estimation and discretization errors and do not permit different samples conditioned on an intermediate state. We present a general method to turn the underlying ODE of such flow models into a family of stochastic differential equations (SDEs) that have the same marginal distributions. This method permits us to derive families of _stochastic samplers_, for fixed (e.g., previously trained) _deterministic_ flow models, that continuously span the spectrum of deterministic and stochastic sampling, given access to the flow field and the score function. Our method provides additional degrees of freedom that help alleviate some of the issues with the deterministic samplers and empirically outperforms them. We demonstrate this empirically on a toy Gaussian setup, as well as on the large scale ImageNet generation task. Further, our family of stochastic samplers provide an additional knob for controlling the diversity of generation, which we qualitatively demonstrate in our experiments.
Primary Area: generative models
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 5002
Loading