Add teaser and badges

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+assets/teaser.webp filter=lfs diff=lfs merge=lfs -text

README.md

@@ -18,8 +18,13 @@ tags:
 
 Compressed FLUX.1 models from **Importance-Aware Low-Rank Distillation of Diffusion Transformers** (ECCV 2026).
 
-[Paper](https://arxiv.org/abs/2609.04646) · [Project page](https://vislearn.github.io/SVDtrunc/) ·
-[Code](https://github.com/vislearn/SVDtrunc)
+[![arXiv](https://img.shields.io/badge/arXiv-2609.04646-b31b1b?logo=arxiv&logoColor=white)](https://arxiv.org/abs/2609.04646)
+[![Project page](https://img.shields.io/badge/Project-Page-0f6b70?logo=googlechrome&logoColor=white)](https://vislearn.github.io/SVDtrunc/)
+[![Code](https://img.shields.io/badge/GitHub-Code-181717?logo=github&logoColor=white)](https://github.com/vislearn/SVDtrunc)
+
+![Two prompts generated by FLUX.1-dev and by SVDtrunc at 68%, 57%, 49% and 39% of its parameters](assets/teaser.webp)
+<sub>Each row uses the same prompt: FLUX.1-dev on the left, then SVDtrunc at 68%, 57%, 49% and 39% of its transformer
+parameters. SVDtrunc-m and SVDtrunc-s are the 68% and 57% models released here.</sub>
 
 | Model               | Base model     | Transformer parameters | File                                             |
 | ------------------- | -------------- | ---------------------- | ------------------------------------------------ |