ELLA: Equip Diffusion Models with LLM for Enhanced Semantic Alignment

By Xi-Wei Hu, Rui Wang, Yixiao Fang, Bin Fu, Pei Cheng, Gang Yu. arXiv.org, 2024

423 citations recorded by the source corpus.

Diffusion models have demonstrated remarkable performance in the domain of text-to-image generation. However, most widely used models still employ CLIP as their text encoder, which constrains their ability to comprehend dense prompts, encompassing multiple objects, detailed attributes, complex relationships, long-text alignment, etc. In this paper, we introduce an Efficient Large Language Model Adapter, termed ELLA, which equips text-to-image diffusion models with powerful Large Language Models (LLM) to enhance text alignment without training of either U-Net or LLM. To seamlessly bridge two pre-trained models, we investigate a range of semantic alignment connector designs and propose a novel module, the Timestep-Aware Semantic Connector (TSC), which dynamically extracts timestep-dependent conditions from LLM. Our approach adapts semantic features at different stages of the denoising process, assisting diffusion models in interpreting lengthy and intricate prompts over sampling timesteps. Additionally, ELLA can be readily incorporated with community models and tools to improve their prompt-following capabilities. To assess text-to-image models in dense prompt following, we introduce Dense Prompt Graph Benchmark (DPG-Bench), a challenging benchmark consisting of 1K dense prompts. Extensive experiments demonstrate the superiority of ELLA in dense prompt following compared to state-of-the-art methods, particularly in multiple object compositions involving diverse attributes and relationships.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/da0d382c7fa981ba185ca633868442b75cb76de6

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Yujun Lin
2. Ligeng Zhu
3. Han Cai
4. Haotian Tang
5. Jingren Zhou
6. Shuai Bai
7. Zhekai Zhang
8. Yichang Zhang

Citations in the corpus

• SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers (2024, positive): “GenEval and DPG-Bench both focus on measuring text-image alignment, with 533 and 1,065 test prompts, respectively.”
• Emu3: Next-Token Prediction is All You Need (2024, methodology): “The results of other methods in the DPG-Benchmark are sourced from [31, 54].”
• Qwen-Image Technical Report (2025, positive): “To assess the model’s general generation performance, we conduct evaluations on four publicly available benchmarks — DPG (Hu et al., 2024b), GenEval (Ghosh et al., 2023), OneIG-Bench (Chang et al., 2025), and TIIF (Wei et al., 2025).”
• Infinity∞: Scaling Bitwise AutoRegressive Modeling for High-Resolution Image Synthesis (2024, neutral): “On DPG [29].”