Saved in:
Bibliographic Details
Main Authors: Tang, Wenpin, Zhou, Fuzhong
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.06279
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908555981881344
author Tang, Wenpin
Zhou, Fuzhong
author_facet Tang, Wenpin
Zhou, Fuzhong
contents This paper aims to develop and provide a rigorous treatment to the problem of entropy regularized fine-tuning in the context of continuous-time diffusion models, which was recently proposed by Uehara et al. (arXiv:2402.15194, 2024). The idea is to use stochastic control for sample generation, where the entropy regularizer is introduced to mitigate reward collapse. We also show how the analysis can be extended to fine-tuning with a general $f$-divergence regularizer. Numerical experiments on large-scale text-to-image models--Stable Diffusion v1.5 are conducted to validate our approach.
format Preprint
id arxiv_https___arxiv_org_abs_2403_06279
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Fine-tuning of diffusion models via stochastic control: entropy regularization and beyond
Tang, Wenpin
Zhou, Fuzhong
Optimization and Control
Machine Learning
This paper aims to develop and provide a rigorous treatment to the problem of entropy regularized fine-tuning in the context of continuous-time diffusion models, which was recently proposed by Uehara et al. (arXiv:2402.15194, 2024). The idea is to use stochastic control for sample generation, where the entropy regularizer is introduced to mitigate reward collapse. We also show how the analysis can be extended to fine-tuning with a general $f$-divergence regularizer. Numerical experiments on large-scale text-to-image models--Stable Diffusion v1.5 are conducted to validate our approach.
title Fine-tuning of diffusion models via stochastic control: entropy regularization and beyond
topic Optimization and Control
Machine Learning
url https://arxiv.org/abs/2403.06279