Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,70 @@
+---
+license: mit
+task_categories:
+- depth-estimation
+- video-classification
+tags:
+- depth
+- normal
+- computer-vision
+- video
+size_categories:
+- 10K<n<100K
+---
+
+# Depth and Normal Videos Dataset
+
+A comprehensive dataset containing **42,498 videos** with depth sequences and surface normals.
+
+## Dataset Summary
+
+- **Total Samples**: 42,498 videos
+- **Groups**: 5 groups (10,000 samples per group)
+- **Source**: Harold328/video_depth
+- **Processing**: Depth Anything V2 + Gradient-based normals
+
+## Structure
+
+```
+Depth_and_Normal_42K/
+├── dataset_metadata.json
+├── group_0000/
+│   ├── videos/   # RGB video sequences
+│   ├── depths/   # Depth video sequences
+│   └── normals/  # Surface normal sequences
+├── group_0001/
+└── ...
+```
+
+## File Naming
+
+Files use 8-digit zero-padded IDs: `00000000.mp4` to `00042497.mp4`
+
+## Video Properties
+
+- Format: MP4
+- Average frames: ~110 per video
+- Frame count range: 83-110 frames
+
+## Usage
+
+```python
+from huggingface_hub import hf_hub_download
+
+# Download a specific video
+video = hf_hub_download(
+    repo_id="Yanbin99/Depth-Normal-Videos-42K",
+    filename="Depth_and_Normal_42K/group_0000/videos/00000000.mp4",
+    repo_type="dataset"
+)
+```
+
+## Applications
+
+- Video depth estimation
+- Temporal consistency learning
+- 3D video reconstruction
+
+## License
+
+MIT License