Upload folder using huggingface_hub

README.md

@@ -0,0 +1,83 @@
+---
+license: apache-2.0
+task_categories:
+  - robotics
+tags:
+  - LeRobot
+  - ur10
+  - robotiq
+  - tactile
+  - cluttered-grasping
+  - imitation-learning
+configs:
+  - config_name: default
+    data_files: data/*/*.parquet
+---
+
+# Cluttered Grasp (UR10 + Robotiq 2F-140 + tactile)
+
+Real-robot teleop demonstrations of cluttered grasping. Episodes are gated on
+`gripper_tcp` height in `base_link` (record while low, pause while high), synced
+to the RealSense RGB stream (~15 Hz), then converted to [LeRobot](https://github.com/huggingface/lerobot) format.
+
+| | |
+|---|---|
+| **Robot** | UR10 + Robotiq 2F-140 (`ur10_robotiq_2f140`) |
+| **FPS** | 15 |
+| **Episodes** | 5 (rebuild grows over collection days) |
+| **Frames** | 844 |
+| **Codebase** | LeRobot dataset v3.0 |
+| **Task** | `cluttered grasping` |
+
+## Dataset structure
+
+```text
+cluttered_grasping/
+├── meta/
+│   ├── info.json          # schema, fps, totals, video codec info
+│   ├── stats.json         # per-feature normalization stats
+│   ├── tasks.parquet      # task prompts
+│   └── episodes/          # per-episode metadata
+├── data/
+│   └── chunk-*/file-*.parquet   # state, action, indices
+└── videos/
+    ├── observation.images.center_cam/chunk-*/file-*.mp4
+    ├── observation.images.tactile_L/chunk-*/file-*.mp4
+    └── observation.images.tactile_R/chunk-*/file-*.mp4
+```
+
+### Features (per frame)
+
+| Key | Type | Shape | Description |
+|-----|------|-------|-------------|
+| `observation.state` | float32 | (14,) | `joint_0..5`, `gripper`, `eef_x/y/z`, `eef_qx/qy/qz/qw` |
+| `action` | float32 | (7,) | Absolute **next** `joint_0..5` + `gripper` (last frame of each episode dropped) |
+| `observation.images.center_cam` | video (AV1) | 720×1280×3 | Wrist/scene RealSense RGB |
+| `observation.images.tactile_L` | video (AV1) | 240×320×3 | Left fingertip tactile (JPEG→RGB) |
+| `observation.images.tactile_R` | video (AV1) | 240×320×3 | Right fingertip tactile |
+| `timestamp` | float32 | (1,) | Time within episode (s) |
+| `frame_index` / `episode_index` / `index` / `task_index` | int64 | (1,) | LeRobot indices |
+
+`observation.state` EE pose is `gripper_tcp` in `base_link`. Depth and F/T remain in the upstream zarr logs and are **not** in this HF export.
+
+### Episode definition
+
+- **Start** when TCP \(z < 0.30\) m  
+- **End** when TCP \(z > 0.31\) m  
+- No recording between episodes (reset / high)
+
+## Load
+
+```python
+from lerobot.datasets.lerobot_dataset import LeRobotDataset
+
+ds = LeRobotDataset("benudavis/cluttered_grasp")
+print(ds)
+frame = ds[0]
+```
+
+## Collection / conversion notes
+
+- ROS bags → one zarr episode per bag (`bag2dataset` under `~/venvs/clutter`, zarr≥3).
+- Offline rebuild: `scripts/zarr_to_lerobot.py` over `~/EDG_Experiment/clutter/**/*.zarr`.
+- Action mode default: `absolute_next` (see `scripts/lerobot_convert/actions.py`).