libero_plus_goal_noise_all_kinova3
收藏资源简介:
该数据集由LeRobot创建,专注于机器人学领域,包含标签如LeRobot、libero、libero-plus、kinova3、multicam和noise。数据集结构包含观测状态(如电机状态、末端执行器状态、关节状态、夹爪状态)、多视角图像观测(包括手腕、前视和侧视图像,均为256x256 RGB视频,帧率20fps)、动作、时间戳、帧索引、episode索引和任务索引。数据集总共有742个episodes、92878帧、10个任务,使用Kinova3型机器人,数据以Parquet格式存储,视频以MP4格式存储,并遵循Apache 2.0许可证。
This dataset was developed by LeRobot, targeting the robotics domain, with tags including LeRobot, libero, libero-plus, kinova3, multicam, and noise. The dataset structure encompasses observation states (e.g., motor states, end-effector states, joint states, and gripper states), multi-view image observations (including wrist, front-facing, and side-facing images, all of which are 256×256 RGB videos at a frame rate of 20 fps), actions, timestamps, frame indices, episode indices, and task indices. In total, the dataset contains 742 episodes, 92,878 frames, and 10 tasks, utilizing the Kinova3 robot. The structured data is stored in Parquet format, while the video files are saved in MP4 format. This dataset is released under the Apache 2.0 license.
数据集概述
- 数据集名称:
max-chr/libero_plus_goal_noise_all_kinova3 - 许可证: Apache-2.0
- 任务类别: 机器人学 (Robotics)
- 标签: LeRobot, libero, libero-plus, kinova3, multicam, noise
- 创建工具: 使用 LeRobot 创建
数据集结构
- 总片段数 (Total Episodes): 742
- 总帧数 (Total Frames): 92,878
- 总任务数 (Total Tasks): 10
- 帧率 (FPS): 20
- 机器人类型: kinova3
- 数据分割: 训练集 (train) 包含 0 到 742 片段
特征 (Features)
数据集包含以下特征,所有特征均为 float32 类型(除另有说明外):
| 特征名称 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
observation.state |
float32 | (8,) | 包含机器人状态:x, y, z, axis_angle1, axis_angle2, axis_angle3, gripper, gripper |
observation.states.ee_state |
float32 | (6,) | 末端执行器状态:x, y, z, axis_angle1, axis_angle2, axis_angle3 |
observation.states.joint_state |
float32 | (7,) | 关节状态:joint_0 到 joint_6 |
observation.states.gripper_state |
float32 | (2,) | 夹爪状态:gripper, gripper |
observation.images.image |
video | (256, 256, 3) | 主摄像头图像 (RGB, AV1 编码, 20 FPS) |
observation.images.wrist_image |
video | (256, 256, 3) | 腕部摄像头图像 (RGB, AV1 编码, 20 FPS) |
observation.images.frontview_image |
video | (256, 256, 3) | 前视野摄像头图像 (RGB, AV1 编码, 20 FPS) |
observation.images.sideview_image |
video | (256, 256, 3) | 侧视野摄像头图像 (RGB, AV1 编码, 20 FPS) |
action |
float32 | (7,) | 动作:x, y, z, axis_angle1, axis_angle2, axis_angle3, gripper |
timestamp |
float32 | (1,) | 时间戳 |
frame_index |
int64 | (1,) | 帧索引 |
episode_index |
int64 | (1,) | 片段索引 |
index |
int64 | (1,) | 索引 |
task_index |
int64 | (1,) | 任务索引 |
数据文件
- 数据以 Parquet 格式存储,路径为
data/*/*.parquet。 - 视频文件以 MP4 格式存储,路径为
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4。 - 数据块大小 (Chunks Size): 1000
- 数据文件大小: 约 100 MB
- 视频文件大小: 约 1e-06 MB
引用
该数据集尚无可引用的 BibTeX 信息。



