FLUX-Text

FLUX-Text: A Simple and Advanced Diffusion Transformer Baseline for Scene Text Editing

Rui Lan^1* Yancheng Bai^{1* ^} Xu Duan¹ Mingxing Li¹ Dongyang Jin¹ Ryan Xu¹ Dong Nie² Lei Sun^{1^ †} Xiangxiang Chu^1†

^*Equal contribution,^{^}Corresponding author,^†Project lead

¹Machine Learning Team, AMAP

²University of North Carolina at Chapel Hill

Research Paper GitHub HuggingFace

Poster Created from FLUX-Text

AMAP FLUX-Text

ALIBABA

Thanks for follow!

Editing of Any Text

FLUX-Text generates text editing results with fidelity and reliability based on the context.

☕

Poster Editing

FLUX-Text can achieve accurate results in poster editing.

☕

Emoticon Editing

FLUX-Text can effectively complete emoticon editing.

☕

Multilingual Editing

FLUX-Text has excellent generalization in multilingual scenarios.

☕

Text Inpainting

FLUX-Text demonstrates excellent capabilities in text inpainting.

☕

Chinese Editing

FLUX-Text demonstrates excellent Chinese editing capabilities in anytext-benchmark.

☕

English Editing

FLUX-Text demonstrates excellent English editing capabilities in anytext-benchmark.

☕

Ethics Concerns

The images used in these demos are sourced from public domains or generated by models, and are intended solely to showcase the capabilities of this research. If you have any concerns, please contact us at lr264907@alibaba-inc.com, and we will promptly remove them.

Acknowledgements

Our work is primarily based on OminiControl, AnyText, Open-Sora, Phantom. We are sincerely grateful for their excellent works.

BibTeX

@article{lan2025flux,
            title={Flux-text: A simple and advanced diffusion transformer baseline for scene text editing},
            author={Lan, Rui and Bai, Yancheng and Duan, Xu and Li, Mingxing and Jin, Dongyang and Xu, Ryan and Sun, Lei and Chu, Xiangxiang},
            journal={arXiv preprint arXiv:2505.03329},
            year={2025}
          }