fix: format dataset structure and quick start code blocks

README.md

@@ -160,14 +160,17 @@ Full taxonomy includes locomotion, manipulation, dexterous hand, tool use, objec
 ---
 
 ## Dataset Structure
+
+```
 chingmu-1000h/
-├── metadata/ ← Index files (CSV, taxonomy)
-├── retargeted/ ← Robot-ready trajectories (gated)
-│ └── g1_joint_trajectory/
-├── raw_bvh/ ← Original BVH (gated)
-├── annotations/ ← Semantic labels
-├── samples/ ← Public preview (no gate)
+├── metadata/          ← Index files (CSV, taxonomy)
+├── retargeted/        ← Robot-ready trajectories (gated)
+│   └── g1_joint_trajectory/
+├── raw_bvh/           ← Original BVH (gated)
+├── annotations/       ← Semantic labels
+├── samples/           ← Public preview (no gate)
 └── LICENSE.md
+```
 
 ---
 
@@ -175,16 +178,20 @@ chingmu-1000h/
 
 ```bash
 pip install huggingface_hub
+```
 
 ```python
 from huggingface_hub import hf_hub_download
+
 repo_id = "ZIHLING/Chingmu-RobotData"
 file_path = hf_hub_download(
-repo_id=repo_id,
-filename="samples/Move_the_box_001.bvh",
-repo_type="dataset",
-local_dir="./robot_samples")
+    repo_id=repo_id,
+    filename="samples/Move_the_box_001.bvh",
+    repo_type="dataset",
+    local_dir="./robot_samples"
+)
 print(f"Downloaded: {file_path}")
+```
 
 For full access, request permission via the **Request access** button, then use `snapshot_download`.
 
@@ -204,8 +211,6 @@ For full access, request permission via the **Request access** button, then use
 
 ## Quality & Limitations
 
-## Quality & Limitations
-
 **Quality controls:** marker swap correction, gap-filling (≤6 frames), foot skating detection, manual review. Flags: `pass`, `warning`, `fail`.
 
 **Accuracy:** joint error <1mm, object pose ±2mm / ±0.5°, temporal sync <1 frame.