SUR-adapter: Enhancing Text-to-Image Pre-trained Diffusion Models with Large Language Models

TL;DR

This paper introduces SUR-adapter, a parameter-efficient method that leverages large language models and a new dataset to improve diffusion models' understanding of narrative prompts for higher-quality text-to-image generation.

Contribution

The paper proposes SUR-adapter, a novel fine-tuning approach that enhances diffusion models' semantic understanding using knowledge distillation from large language models and a new annotated dataset.

Findings

Improved image quality with narrative prompts

Enhanced semantic understanding in diffusion models

Effective knowledge transfer from LLMs to diffusion models

Abstract

Diffusion models, which have emerged to become popular text-to-image generation models, can produce high-quality and content-rich images guided by textual prompts. However, there are limitations to semantic understanding and commonsense reasoning in existing models when the input prompts are concise narrative, resulting in low-quality image generation. To improve the capacities for narrative prompts, we propose a simple-yet-effective parameter-efficient fine-tuning approach called the Semantic Understanding and Reasoning adapter (SUR-adapter) for pre-trained diffusion models. To reach this goal, we first collect and annotate a new dataset SURD which consists of more than 57,000 semantically corrected multi-modal samples. Each sample contains a simple narrative prompt, a complex keyword-based prompt, and a high-quality image. Then, we align the semantic representation of narrative prompts to the complex prompts and transfer knowledge of large language models (LLMs) to our SUR-adapter via knowledge distillation so that it can acquire the powerful semantic understanding and reasoning capabilities to build a high-quality textual semantic representation for text-to-image generation. We conduct experiments by integrating multiple LLMs and popular pre-trained diffusion models to show the effectiveness of our approach in enabling diffusion models to understand and reason concise natural language without image quality degradation. Our approach can make text-to-image diffusion models easier to use with better user experience, which demonstrates our approach has the potential for further advancing the development of user-friendly text-to-image generation models by bridging the semantic gap between simple narrative prompts and complex keyword-based prompts. The code is released at https://github.com/Qrange-group/SUR-adapter.