Header links: code, weights, dataset, project page

README.md

@@ -19,7 +19,9 @@ negative disparity. Image models are fine-tuned on translated SceneFlow only;
 video models on translated SceneFlow + Dynamic Replica, matching the data of
 their released baselines. The ZDPShift benchmark itself is never trained on.
 
-Weights: [shijianjian/ZDPShift](https://huggingface.co/shijianjian/ZDPShift).
+Code: [github.com/shijianjian/ZDPShift](https://github.com/shijianjian/ZDPShift) ·
+Weights: [shijianjian/ZDPShift](https://huggingface.co/shijianjian/ZDPShift) ·
+Project page: [shijianjian.github.io/ZDPShift](https://shijianjian.github.io/ZDPShift/).
 Rendered examples come from Blender Studio open movies (CC-BY) at the largest
 ZDP shift, where most of the scene sits behind the screen plane. The film
 examples (Jurassic World, Avatar, Galapagos, Kung Fu Panda) are delivered

app.py

@@ -64,6 +64,10 @@ def build_examples():
 with gr.Blocks(title="Unmuzzling Stereo") as demo:
     gr.Markdown(
         "# Unmuzzling Stereo — signed disparity for content behind the screen\n"
+        "[Code](https://github.com/shijianjian/ZDPShift) · "
+        "[Weights](https://huggingface.co/shijianjian/ZDPShift) · "
+        "[Dataset](https://huggingface.co/datasets/shijianjian/ZDPShift) · "
+        "[Project page](https://shijianjian.github.io/ZDPShift/)\n\n"
         "Released stereo matchers cannot report negative disparity. The corrected "
         "models — fine-tuned on horizontally translated SceneFlow only — recover "
         "it with no architectural change. Red = in front of the screen plane, "