poolvarine/robomimic-mh-square-image-dense

---
task_categories:
- robotics
dataset_info:
features:
- name: action
- name: next.done
- name: next.reward
- name: observation.state
- name: timestamp
- name: frame_index
- name: episode_index
- name: index
- name: task_index
splits:
- name: train
num_examples: 300
---
# robomimic-mh-square-image-dense

This dataset was created using [LeRobot](https://github.com/huggingface/lerobot).

## Dataset Description

A robotics demonstration dataset for training imitation learning and reinforcement learning policies.

### Dataset Statistics

- **Total Episodes:** 300
- **Total Frames:** 80731
- **Total Tasks:** 1
- **FPS:** 20
- **Robot Type:** robomimic
- **Codebase Version:** v2.1

## Dataset Structure

The dataset directory contains:

- `meta/`: Metadata files
- `info.json`: Dataset configuration
- `episodes.jsonl`: Episode metadata
- `episodes_stats.jsonl`: Episode statistics
- `tasks.jsonl`: Task information

- `data/`: Parquet data files
- `train-00000-of-00001.parquet`: Consolidated dataset

- `videos/`: Video files (if available)
- `episode_*.mp4`: Episode videos

## Dataset Metadata

### info.json

Contains comprehensive dataset metadata and configuration:

```json
{
"codebase_version": "v2.1",
"robot_type": "robomimic",
"total_episodes": 300,
"total_frames": 80731,
"total_tasks": 1,
"total_videos": 22,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:11"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4",
"features": {
"action": {
"dtype": "float32",
"shape": [
7
],
"names": [
"eef_delta_pos_x",
"eef_delta_pos_y",
"eef_delta_pos_z",
"eef_delta_rot_rx",
"eef_delta_rot_ry",
"eef_delta_rot_rz",
"gripper_action"
]
},
"next.done": {
"dtype": "bool",
"shape": [
1
],
"names": [
"done"
]
},
"next.reward": {
"dtype": "float32",
"shape": [
1
],
"names": [
"reward"
]
},
"observation.state": {
"dtype": "float32",
"shape": [
9
],
"names": [
"robot0_eef_pos_0",
"robot0_eef_pos_1",
"robot0_eef_pos_2",
"robot0_eef_quat_0",
"robot0_eef_quat_1",
"robot0_eef_quat_2",
"robot0_eef_quat_3",
"robot0_gripper_qpos_0",
"robot0_gripper_qpos_1"
]
},
"observation.images.agentview": {
"dtype": "video",
"shape": [
256,
256,
3
],
"names": [
"height",
"width",
"channel"
],
"info": {
"video.height": 256,
"video.width": 256,
"video.codec": "av1",
"video.pix_fmt": "yuv420p",
"video.is_depth_map": false,
"video.fps": 20,
"video.channels": 3,
"has_audio": false
}
},
"observation.images.robot0_eye_in_hand": {
"dtype": "video",
"shape": [
256,
256,
3
],
"names": [
"height",
"width",
"channel"
],
"info": {
"video.height": 256,
"video.width": 256,
"video.codec": "av1",
"video.pix_fmt": "yuv420p",
"video.is_depth_map": false,
"video.fps": 20,
"video.channels": 3,
"has_audio": false
}
},
"timestamp": {
"dtype": "float32",
"shape": [
1
],
"names": null
},
"frame_index": {
"dtype": "int64",
"shape": [
1
],
"names": null
},
"episode_index": {
"dtype": "int64",
"shape": [
1
],
"names": null
},
"index": {
"dtype": "int64",
"shape": [
1
],
"names": null
},
"task_index": {
"dtype": "int64",
"shape": [
1
],
"names": null
}
}
}
```

## Features

- **action** (`float32`) - Shape: [7] - Components: eef_delta_pos_x, eef_delta_pos_y, eef_delta_pos_z ... (+4 more)
- **next.done** (`bool`) - Shape: [1] - Components: done
- **next.reward** (`float32`) - Shape: [1] - Components: reward
- **observation.state** (`float32`) - Shape: [9] - Components: robot0_eef_pos_0, robot0_eef_pos_1, robot0_eef_pos_2 ... (+6 more)
- **observation.images.agentview** (`video`) - Shape: [256, 256, 3] - Components: height, width, channel
- **observation.images.robot0_eye_in_hand** (`video`) - Shape: [256, 256, 3] - Components: height, width, channel
- **timestamp** (`float32`) - Shape: [1]
- **frame_index** (`int64`) - Shape: [1]
- **episode_index** (`int64`) - Shape: [1]
- **index** (`int64`) - Shape: [1]
- **task_index** (`int64`) - Shape: [1]
## Usage

To load this dataset with LeRobot:

```python
from lerobot.common.datasets import LeRobotDataset
dataset = LeRobotDataset(repo_id="poolvarine/robomimic-mh-square-image-dense")
```

Or with the Hugging Face Datasets library:

```python
from datasets import load_dataset
dataset = load_dataset("poolvarine/robomimic-mh-square-image-dense")
```

## Citation

If you use this dataset, please cite LeRobot:

```bibtex
@software{ lerobot2024,
author = { The HuggingFace Team },
title = { LeRobot: A Python Library for Robot Learning },
url = { https://github.com/huggingface/lerobot },
year = { 2024 }
}
```

## License

[More Information Needed]

## Dataset Card Contact

[More Information Needed]