metaworld_mt50_push_v2_image

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

## Dataset Description

We reposition the camera and flip the rendered images as follow:
```env.model.cam_pos[2] = [0.75, 0.075, 0.7]
```

- **Homepage:** [More Information Needed]
- **Paper:** [More Information Needed]
- **License:** apache-2.0

## Dataset Structure

[meta/info.json](meta/info.json):
```json
{
"codebase_version": "v2.0",
"robot_type": "metaworld",
"total_episodes": 100,
"total_frames": 6125,
"total_tasks": 1,
"total_videos": 0,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 80,
"splits": {
"train": "0:100"
},
"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": {
"observation.state": {
"dtype": "float32",
"shape": [
4
],
"names": {
"axes": null
}
},
"action": {
"dtype": "float32",
"shape": [
4
],
"names": {
"axes": [
"x",
"y",
"z",
"gripper"
]
}
},
"next.reward": {
"dtype": "float32",
"shape": [
1
],
"names": null
},
"next.success": {
"dtype": "bool",
"shape": [
1
],
"names": null
},
"observation.environment_state": {
"dtype": "float32",
"shape": [
39
],
"names": [
"keypoints"
]
},
"observation.image": {
"dtype": "image",
"shape": [
3,
480,
480
],
"names": [
"channels",
"height",
"width"
]
},
"task_id": {
"dtype": "int16",
"shape": [
1
],
"names": null
},
"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
}
}
}
```

## Citation

**BibTeX:**

```bibtex
[More Information Needed]
```

本数据集基于[LeRobot](https://github.com/huggingface/lerobot)构建。

## 数据集描述
我们对相机位姿进行调整并翻转渲染图像，具体操作如下：
python
env.model.cam_pos[2] = [0.75, 0.075, 0.7]

- **项目主页**：[需补充更多信息]
- **相关论文**：[需补充更多信息]
- **授权协议**：Apache-2.0

## 数据集结构
`meta/info.json` 文件内容如下：
json
{
"代码库版本": "v2.0",
"机器人类型": "metaworld",
"总回合数": 100,
"总帧数": 6125,
"总任务数": 1,
"总视频数": 0,
"总分块数": 1,
"分块大小": 1000,
"帧率": 80,
"数据集划分": {
"训练集": "0:100"
},
"数据存储路径": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"视频存储路径": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4",
"数据特征": {
"观测.状态": {
"数据类型": "float32",
"形状": [
4
],
"维度命名": {
"轴": null
}
},
"动作": {
"数据类型": "float32",
"形状": [
4
],
"维度命名": {
"轴": [
"x",
"y",
"z",
"夹爪"
]
}
},
"下一时刻奖励": {
"数据类型": "float32",
"形状": [
1
],
"维度命名": null
},
"下一时刻任务成功标识": {
"数据类型": "bool",
"形状": [
1
],
"维度命名": null
},
"观测.环境状态": {
"数据类型": "float32",
"形状": [
39
],
"维度命名": [
"关键点"
]
},
"观测.图像": {
"数据类型": "图像",
"形状": [
3,
480,
480
],
"维度命名": [
"通道数",
"高度",
"宽度"
]
},
"任务ID": {
"数据类型": "int16",
"形状": [
1
],
"维度命名": null
},
"时间戳": {
"数据类型": "float32",
"形状": [
1
],
"维度命名": null
},
"帧索引": {
"数据类型": "int64",
"形状": [
1
],
"维度命名": null
},
"回合索引": {
"数据类型": "int64",
"形状": [
1
],
"维度命名": null
},
"全局索引": {
"数据类型": "int64",
"形状": [
1
],
"维度命名": null
},
"任务索引": {
"数据类型": "int64",
"形状": [
1
],
"维度命名": null
}
}
}

## 引用
**BibTeX格式**：
bibtex
[需补充更多信息]

搜集汇总

背景与挑战

背景概述

该数据集由LeRobot创建，包含100个训练片段和6125帧数据，专门用于机器人推任务。数据集以Parquet和视频格式存储，包含状态、动作、奖励、图像等多维度特征，总大小为1.38GB。