| license: mit | |
| task_categories: | |
| - robotics | |
| tags: | |
| - LeRobot | |
| - SO-100 | |
| - robotics | |
| - towel-folding | |
| # Trimmed Towel Folding Dataset | |
| This dataset contains towel folding demonstrations trimmed to the first fold. | |
| ## Dataset Details | |
| - **Original Dataset**: [arsenxeno/augmented-setup](https://huggingface.co/datasets/arsenxeno/augmented-setup) | |
| - **Task**: Towel folding (first fold only) | |
| - **Episodes**: 1 | |
| - **Video Feeds**: 3 | |
| - **Codebase Version**: v2.0 | |
| - **Robot Type**: SO-100 | |
| ## Statistics | |
| | Metric | Value | | |
| |--------|-------| | |
| | Total Episodes | 1 | | |
| | Total Video Feeds | 3 | | |
| | Average Original Length | 709 frames | | |
| | Average Trimmed Length | 338 frames | | |
| | Average Fold Time | 22.5 seconds | | |
| ## Usage with LeRobot | |
| ```python | |
| from lerobot.datasets import make_dataset | |
| # Load the dataset | |
| dataset = make_dataset("xenorobotics/augmented-setup-2") | |
| # Access different video streams | |
| front_videos = dataset["observation.images.front"] | |
| # If available: left_videos = dataset["observation.images.left"] | |
| ``` | |
| ## Dataset Structure | |
| ``` | |
| . | |
| ├── data/ | |
| │ └── chunk-000/ | |
| │ ├── episode_000000.parquet | |
| │ ├── episode_000001.parquet | |
| │ └── ... | |
| ├── videos/ | |
| │ └── chunk-000/ | |
| │ ├── observation.images.front/ | |
| │ │ ├── episode_000000.mp4 | |
| │ │ └── ... | |
| │ └── [other video feeds]/ | |
| ├── meta/ | |
| │ ├── info.json | |
| │ ├── episodes.jsonl | |
| │ ├── tasks.jsonl | |
| │ └── stats.json | |
| └── README.md | |
| ``` | |