pith. sign in

arxiv: 2304.00186 · v5 · pith:PHFCAWXInew · submitted 2023-04-01 · 💻 cs.CV · cs.AI

Subject-driven Text-to-Image Generation via Apprenticeship Learning

classification 💻 cs.CV cs.AI
keywords subjectsutiexpertmodelmodelssubject-specificlearningtext-to-image
0
0 comments X
read the original abstract

Recent text-to-image generation models like DreamBooth have made remarkable progress in generating highly customized images of a target subject, by fine-tuning an ``expert model'' for a given subject from a few examples. However, this process is expensive, since a new expert model must be learned for each subject. In this paper, we present SuTI, a Subject-driven Text-to-Image generator that replaces subject-specific fine tuning with in-context learning. Given a few demonstrations of a new subject, SuTI can instantly generate novel renditions of the subject in different scenes, without any subject-specific optimization. SuTI is powered by apprenticeship learning, where a single apprentice model is learned from data generated by a massive number of subject-specific expert models. Specifically, we mine millions of image clusters from the Internet, each centered around a specific visual subject. We adopt these clusters to train a massive number of expert models, each specializing in a different subject. The apprentice model SuTI then learns to imitate the behavior of these fine-tuned experts. SuTI can generate high-quality and customized subject-specific images 20x faster than optimization-based SoTA methods. On the challenging DreamBench and DreamBench-v2, our human evaluation shows that SuTI significantly outperforms existing models like InstructPix2Pix, Textual Inversion, Imagic, Prompt2Prompt, Re-Imagen and DreamBooth, especially on the subject and text alignment aspects.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. TextBoost: Boosting Text Encoder for Personalized Text-to-Image Generation

    cs.CV 2024-09 unverdicted novelty 4.0

    TextBoost is a one-shot personalization technique that selectively fine-tunes the text encoder of diffusion models using causality-preserving adaptation and lightweight adapters to reduce parameters and storage.