BriefGPT.xyz
May, 2023
流匹配方法的误差界
Error Bounds for Flow Matching Methods
HTML
PDF
Joe Benton, George Deligiannidis, Arnaud Doucet
TL;DR
本文提出了一种使用完全确定性采样的流匹配过程,该过程利用概率流ODE方法和去噪扩散隐式模型,为基于ODE的生成模型提供了误差界限。
Abstract
Score-based
generative models
are a popular class of generative modelling techniques relying on
stochastic differential equations
(SDE). From their inception, it was realized that it was also possible to perform
→