Improve model card

README.md

@@ -1,3 +1,27 @@
----
-license: apache-2.0
----
+---
+license: apache-2.0
+pipeline_tag: image-segmentation
+library_name: transformers
+---
+
+# WOW-Seg: A Word-free Open World Segmentation Model
+
+WOW-Seg is a Word-free Open World Segmentation model for segmenting and recognizing objects from open-set categories. It introduces a novel visual prompt module, **Mask2Token**, which transforms image masks into visual tokens and ensures their alignment with the VLLM feature space. Moreover, it introduces the **Cascade Attention Mask** to decouple information across different instances, mitigating inter-instance interference and improving model performance.
+
+- **Paper:** [WOW-Seg: A Word-free Open World Segmentation Model](https://huggingface.co/papers/2605.16903)
+- **Repository:** [https://github.com/AAwcAA/WOW-Seg-Meta](https://github.com/AAwcAA/WOW-Seg-Meta)
+- **Dataset:** [RR-7K Dataset](https://huggingface.co/datasets/AAwcAA/RR-7K)
+
+## Citation
+
+If you find WOW-Seg useful for your research, please use the following BibTeX entry:
+
+```bibtex
+@inproceedings{li2026wowseg,
+      title={{WOW}-Seg: A Word-free Open World Segmentation Model},
+      author={Danyang Li and Tianhao Wu and Bin Lin and Zhenyuan Chen and Yang Zhang and Yuxuan Li and Ming-Ming Cheng and Xiang Li},
+      booktitle={The Fourteenth International Conference on Learning Representations},
+      year={2026},
+      url={https://openreview.net/forum?id=AyJPSnE1bq}
+}
+```