Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,48 @@
+---
+license: apache-2.0
+tags:
+- robotics
+- manipulation
+- libero
+- multimodal
+- task-images
+task_categories:
+- robotics
+- computer-vision
+- multimodal
+---
+
+# test_lerobot
+
+This dataset contains robotic manipulation data extracted from LeRobot datasets.
+
+## Dataset Structure
+
+- `data/`: Episode data in parquet format
+- `meta/`: Metadata including episodes, tasks, and statistics
+- `task_images/`: Task-related multimodal images (reference, goal, instruction images)
+
+## Task Images
+
+This dataset includes multimodal task images:
+- **Reference Images**: Visual examples of each task
+- **Goal Images**: Target states extracted from successful episodes  
+- **Instruction Images**: Visual task descriptions
+
+## Usage
+
+```python
+from lerobot.datasets.lerobot_dataset import LeRobotDataset
+
+# Load the dataset
+dataset = LeRobotDataset("vgyuan/test_lerobot")
+
+# Access task images metadata
+import json
+with open("task_images/task_images_metadata.json") as f:
+    task_images_info = json.load(f)
+```
+
+## Original Source
+
+This dataset is extracted from the original LeRobot dataset for specific robotic manipulation tasks.