| --- |
| pipeline_tag: image-to-image |
| tags: |
| - flux |
| - product-poster-generation |
| - text-rendering |
| - image-generation |
| - cvpr-2026 |
| language: |
| - en |
| - zh |
| --- |
| |
| # Flux-SimplePoster |
|
|
| Official SIMPLEPOSTER inference checkpoint for **SIMPLEPOSTER: A Simple Baseline for Product Poster Generation** (CVPR 2026). |
|
|
| - [Paper](https://arxiv.org/abs/2605.08784) |
| - [Official GitHub](https://github.com/Alibaba-VELLDEPTH/SIMPLEPOSTER) |
| - [Internal inference project](https://code.alibaba-inc.com/Alibaba-Yufeng/simple_poster) |
|
|
| ## Model file |
|
|
| | File | Description | |
| |---|---| |
| | `dit.pth` | SIMPLEPOSTER DiT checkpoint (`checkpoint_0041012`) | |
|
|
| The checkpoint is used together with the FLUX.1-dev autoencoder and Qwen2.5-VL-7B-Instruct. Those dependency weights are not included in this repository and remain subject to their respective licenses and terms. |
|
|
| ## Download |
|
|
| ```bash |
| hf download Alibaba-VELLDEPTH/Flux-SimplePoster dit.pth --local-dir models |
| ``` |
|
|
| For inference code and setup instructions, see the [official GitHub repository](https://github.com/Alibaba-VELLDEPTH/SIMPLEPOSTER). |
|
|
| ## Recommended inference settings |
|
|
| - Resolution: 1024 × 1024 |
| - Sampling steps: 50 |
| - `guidance_scale`: 4.5 |
| - `guidance`: 30 |
| - Seed: 100 |
|
|
| ## Citation |
|
|
| ```bibtex |
| @article{simpleposter2026, |
| title={SIMPLEPOSTER: A Simple Baseline for Product Poster Generation}, |
| author={Cui, Benlei and Zeng, Fangao and Jiang, Weitao and Zhai, Yuwen and Hong, Haiwen and Huang, Longtao and Xue, Hui and Shang, Wenxiang and Huang, Pipei}, |
| journal={arXiv preprint arXiv:2605.08784}, |
| year={2026} |
| } |
| ``` |
|
|
| ## Intended use and limitations |
|
|
| This checkpoint is intended for research on product-poster generation. Generated text can contain spelling, typography, or layout errors. Users are responsible for reviewing generated content and complying with applicable laws, third-party rights, and the terms of all dependency models. |
|
|
