Add model card

README.md

@@ -1,3 +1,15 @@
 ---
 license: mit
+pipeline_tag: depth-estimation
 ---
+
+# MaSS
+
+This repository contains the MaSS model presented in [Learning to Discriminate Scene Structures Makes Self-Supervised Depth Learning Scalable](https://huggingface.co/papers/2610.23600).
+
+MaSS makes self-supervised monocular depth learning scalable to diverse scenes. It separates contextual and structural representations, helping a single depth model learn across indoor and outdoor environments, dynamic scenes, adverse weather, and beyond.
+
+- **Project page:** https://mias.group/MaSS
+- **Code:** https://github.com/mengtanZ/MaSS
+
+For details about training, evaluation, checkpoints, and usage, please refer to the [GitHub repository](https://github.com/mengtanZ/MaSS).