eval_mix_nooutlier_15k_blue_r2
收藏Hugging Face2026-05-18 更新2026-05-18 收录
下载链接:
https://huggingface.co/datasets/robot-learning-group47/eval_mix_nooutlier_15k_blue_r2
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是一个用于机器人学习的数据集,使用LeRobot工具创建。数据集包含一个episode,共478帧,涉及一个任务,数据以parquet文件格式存储。特征包括动作(6维浮点数组,表示肩部、肘部、腕部和夹爪的位置)、观测状态(与动作相同的6维关节位置)、来自camera1的图像观测(视频格式,分辨率为480x640,帧率30fps,编码为av1),以及时间戳、帧索引、episode索引等元数据。机器人类型为so_follower,适用于机器人控制与模仿学习任务。
This dataset is designed for robotic learning applications and was developed using the LeRobot toolkit. It comprises one episode containing a total of 478 frames for a single task, and is stored in Parquet file format. The dataset features include: actions (a 6-dimensional floating-point array representing the positions of the shoulder, elbow, wrist, and gripper), observation states (6-dimensional joint positions matching those of the actions), image observations from Camera 1 (video format with a resolution of 480x640, 30 frames per second, encoded in AV1), as well as metadata such as timestamps, frame indices, and episode indices. The robot type is so_follower, which is suitable for robotic control and imitation learning tasks.
创建时间:
2026-05-18
原始信息汇总
数据集概述
数据集名称:eval_mix_nooutlier_15k_blue_r2
许可证:Apache-2.0
任务类别:机器人学(Robotics)
标签:LeRobot
数据集来源与创建
数据集统计信息
- 总片段数 (Episodes):1
- 总帧数 (Frames):478
- 总任务数:1
- 机器人类型:
so_follower - 帧率 (FPS):30
- 数据文件大小:约 100 MB
- 视频文件大小:约 200 MB
- 训练/测试划分:仅训练集(
train: 0:1)
数据结构
数据集以分块形式存储,包含 Parquet 文件(数据)和 MP4 视频文件(观测图像)。
特征 (Features):
| 特征名 | 数据类型 | 说明 |
|---|---|---|
action |
float32, shape [6] | 机器人动作指令,包含 6 个关节位置(肩部旋转、肩部升降、肘部弯曲、腕部弯曲、腕部旋转、夹爪) |
observation.state |
float32, shape [6] | 机器人观测状态,同样为 6 个关节位置 |
observation.images.camera1 |
video, shape (480, 640, 3) | 来自摄像机1的 RGB 观测图像,分辨率 480x640,帧率 30fps,编码为 AV1 |
timestamp |
float32, shape [1] | 时间戳 |
frame_index |
int64, shape [1] | 帧索引 |
episode_index |
int64, shape [1] | 片段索引 |
index |
int64, shape [1] | 全局索引 |
task_index |
int64, shape [1] | 任务索引 |
数据路径
- 数据文件:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频文件:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4
引用信息
目前暂无可用的 BibTeX 引用。



