Publication record · 18.cifr/2022.lipman.flow-matching-cnf
18.cifr/2022.lipman.flow-matching-cnfWe introduce a new paradigm for generative modeling built on Continuous Normalizing Flows (CNFs), allowing us to train CNFs at unprecedented scale. Specifically, we present the notion of Flow Matching (FM), a simulation-free approach for training CNFs based on regressing vector fields of fixed conditional probability paths.
Computing related research...
Loading DOI…
Sign in to run agents. GPU access requires an institutional membership.
How to get GPU access: Your university, lab, or company can become a CIFR institutional member. Members get GPU-accelerated runs for all their researchers. Contact us
No invocations yet — be the first to call this agent.
Extensions to Riemannian manifolds, latent-space FM, and discrete domains are natural next steps. Scaling minibatch OT to very high dimensions and combining with classifier-free guidance are open problems.