stackblocks_recap_all_for_vf_v2
收藏资源简介:
该数据集使用LeRobot创建,专注于机器人学任务,特别是堆叠积木的演示数据。数据集包含机器人动作(如肩部平移、肩部升降、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置)和观测状态(包括关节位置以及来自顶部和腕部摄像头的视频图像,分辨率为480x640,帧率为30fps)。此外,还包括时间戳、帧索引、回合索引、任务索引、终端状态、蒙特卡洛回报、干预标志和下一步成功标志等元数据。数据以Parquet格式组织,总共有267个回合、216,929帧和1个任务,总数据大小约为300MB(数据文件100MB,视频文件200MB)。数据集适用于机器人控制、强化学习和计算机视觉研究。
This dataset is created using LeRobot and focuses on robotics tasks, specifically demonstration data for stacking blocks. It includes robot actions (such as shoulder translation, shoulder lift, elbow bend, wrist bend, wrist rotation, and gripper position) and observation states (including joint positions and video images from top and wrist cameras, with a resolution of 480x640 and a frame rate of 30fps). Additionally, it contains metadata such as timestamps, frame index, episode index, task index, terminal state, Monte Carlo return, intervention flag, and next success flag. The data is organized in Parquet format, with a total of 267 episodes, 216,929 frames, and 1 task, and a total data size of approximately 300MB (100MB for data files and 200MB for video files). The dataset is suitable for research in robot control, reinforcement learning, and computer vision.
数据集概览
- 数据集名称:stackblocks_recap_all_for_vf_v2
- Hugging Face 地址:https://huggingface.co/datasets/lilkm/stackblocks_recap_all_for_vf_v2
- 许可证:Apache-2.0
- 任务类别:机器人学 (Robotics)
- 框架:基于 LeRobot 创建
数据集结构
版本信息:
- 代码库版本:v3.0
- 帧率 (FPS):30
特征 (Features):
| 特征名 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
action |
float32 | (6,) | 6维动作:肩部水平旋转、肩部升降、肘部弯曲、腕部弯曲、腕部旋转、夹爪位置 |
observation.state |
float32 | (6,) | 6维观测状态,同动作空间定义 |
observation.images.top |
视频 (AV1编码) | (480, 640, 3) | 顶部摄像头 RGB 图像,分辨率 480x640 |
observation.images.wrist |
视频 (AV1编码) | (480, 640, 3) | 腕部摄像头 RGB 图像,分辨率 480x640 |
timestamp |
float32 | (1,) | 时间戳 |
frame_index |
int64 | (1,) | 帧索引 |
episode_index |
int64 | (1,) | 片段索引 |
index |
int64 | (1,) | 全局索引 |
task_index |
int64 | (1,) | 任务索引 |
is_terminal |
bool | (1,) | 是否为终止帧 |
mc_return |
float32 | (1,) | Monte Carlo 回报值 |
intervention |
bool | (1,) | 是否有人工干预 |
next.success |
bool | (1,) | 下一个状态是否成功 |
数据规模:
- 总片段数:267
- 总帧数:216,929
- 任务数量:1
- 数据文件大小:约 100 MB
- 视频文件大小:约 200 MB
- 数据文件路径:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频文件路径:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4
机器人类型:so_follower
数据集划分:
- 训练集:片段索引 0 至 266(共 267 个片段)



