Update README.md

README.md

@@ -22,7 +22,7 @@ While recent progress in AI and computer vision has been remarkable, there remai
 Each sub-dataset (scene) contains:
 
 - `images/`: Rendered images under different camera views and backgrounds.
-- `metadata.csv`: Instance-level annotations including object attributes and positions.
+- `tabular.csv`: Instance-level annotations including object attributes in causal graph.
 <!-- - `causal_graph.json`: Ground-truth causal structure (as adjacency matrix or graph).
 - `view_info.json`: Camera/viewpoint metadata.
 - `split.json`: Recommended train/val/test splits for benchmarking. -->