English

ERNIE-Image Technical Report

Computer Vision and Pattern Recognition 2026-05-26 v1 Machine Learning

Abstract

We introduce ERNIE-Image, an open-source text-to-image generation model built upon an 8B single-stream DiT architecture. ERNIE-Image aims to bridge the gap between current open-source models and leading closed-source systems through more effective mining of large-scale pre-training data and improved supervision quality throughout training. During pre-training, we adopt a bottom-up data construction pipeline that combines fine-grained image categorization, rich caption annotation, aesthetic assessment, and hierarchical sampling. This strategy reduces data noise while preserving long-tail concepts and detailed real-world knowledge, providing a stronger foundation for complex generation tasks. In the post-training stage, we use a top-down data construction pipeline for high-demand scenarios, diversify prompt annotations to better match real user inputs, and apply a stabilized DPO strategy to align the model with human aesthetic preferences. We further train ERNIE-Image-Turbo for efficient 8-NFE generation and propose MT-DMD to mitigate capability drift during distillation. To make the model easier to use in practical scenarios, we equip it with a lightweight Prompt Enhancer that expands concise user intents into structured visual descriptions. In addition, we develop ERNIE-Image-Aes, an industrial-grade aesthetic model, together with ERNIE-Image-Aes-1K, a human-annotated benchmark for realistic aesthetic evaluation. Extensive qualitative and quantitative experiments show that ERNIE-Image achieves leading performance among open-source models and approaches top-tier commercial models in instruction following, text rendering, and aesthetic quality. We release the trained models and aesthetic resources to facilitate further academic research and technical progress in the AIGC community.

Keywords

Cite

@article{arxiv.2605.25347,
  title  = {ERNIE-Image Technical Report},
  author = {Jiaxiang Liu and Zhida Feng and Pengyu Zou and Zhenyu Qian and Tianrui Zhu and Jun Xia and Yuehu Dong and Yanzheng Lin and Honglin Xiong and Anqi Chen and Yunpeng Ding and Jinghui Duan and Lin Gao and Chao Han and Tiechao He and Jiakang Hu and Ranjun Hua and Xueming Jiang and Qingli Kong and Yuting Lei and Tianyu Li and Yunlin Liu and Changling Liu and Yaxin Liu and Yi Liu and Xuguang Liu and Xiaolong Ma and Yan Pan and Yiran Ren and Nan Sheng and Yu Sun and Siyang Sun and Yixiang Tu and Yang Wan and Huanai Wang and Siqi Wang and Yang Wu and Youzhi Yang and Xiaowen Yang and Jianwen Yang and Yehua Yang and Quanwen Zhang and Xinmin Zhang and Haoxin Zhang and Xiang Zhang and Jun Zhang and Qian Zhang and Qiao Zhao and Qi Zhou},
  journal= {arXiv preprint arXiv:2605.25347},
  year   = {2026}
}