demo_src_mug_cleanup_task_D1_robot_IIWA_gripper_Robotiq85Gripper
收藏资源简介：
该数据集是一个用于机器人模仿学习和跨体现研究的机器人数据集，由LeRobot工具创建。它包含100个episodes，总计35,073帧，专注于单个任务（可能涉及清洁或操作场景，如基于标签和上下文推断的mug_cleanup）。数据以parquet文件格式存储，并包括视频文件（总计200 MB）。特征包括动作（7维浮点数，表示位移、旋转和夹爪控制）、观测（如状态、末端执行器位置和姿态、夹爪位置、关节位置，以及组合特征如eef_gripper和joint_gripper），以及来自agentview和wrist视角的RGB图像观测（256x256分辨率，20 FPS）。此外，还包含时间戳、框架索引、episode索引等元数据。数据集适用于机器人控制、行为克隆和强化学习研究，支持跨不同机器人体现的泛化。
This dataset is a robotic dataset for robot imitation learning and cross-embodiment research, developed using the LeRobot toolkit. It contains 100 episodes and a total of 35,073 frames, focusing on a single task that may involve cleaning or manipulation scenarios, such as mug_cleanup as inferred from labels and context. The data is stored in Parquet file format and includes video files totaling 200 MB. The features include actions (7-dimensional floating-point numbers representing displacement, rotation and gripper control), observations such as states, end-effector position and pose, gripper position, joint positions, and combined features like eef_gripper and joint_gripper, as well as RGB image observations from agentview and wrist viewpoints with a resolution of 256×256 and 20 FPS. Additionally, it includes metadata such as timestamps, frame indices, episode indices and the like. This dataset is suitable for research on robot control, behavioral cloning and reinforcement learning, and supports generalization across different robot embodiments.
数据集概述
- 数据集名称:
demo_src_mug_cleanup_task_D1_robot_IIWA_gripper_Robotiq85Gripper - 许可证: Apache-2.0
- 任务类别: 机器人学 (Robotics)
- 标签: LeRobot, MimicGen, Cross-Embodiment
- 创建工具: LeRobot
数据集规模与结构
| 属性 | 数值 |
|---|---|
| 总片段数 (Episodes) | 100 |
| 总帧数 (Frames) | 35,073 |
| 总任务数 | 1 |
| 数据文件大小 | 100 MB |
| 视频文件大小 | 200 MB |
| 帧率 (FPS) | 20 |
| 分块大小 (Chunks) | 1,000 |
| 数据集划分 | 训练集: 片段 0~99 (100%) |
数据特征
数据集包含以下特征字段，均为数值型或视频型数据：
| 特征名称 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
action |
float32 | (7,) | 机器人动作：dx, dy, dz, drx, dry, drz, gripper |
observation.state |
float32 | (2,) | 状态观测：zero0, zero1 |
observation.eef |
float32 | (7,) | 末端执行器位姿：位置 (x, y, z) + 四元数 (x, y, z, w) |
observation.gripper |
float32 | (6,) | 夹爪关节位置：gripper_qpos_0 ~ gripper_qpos_5 |
observation.joint |
float32 | (7,) | 关节位置：joint_pos_0 ~ joint_pos_6 |
observation.eef_gripper |
float32 | (13,) | 末端执行器位姿与夹爪关节位置拼接 |
observation.joint_gripper |
float32 | (13,) | 关节位置与夹爪关节位置拼接 |
observation.images.agentview |
video | (256, 256, 3) | 主视角 RGB 视频，AV1 编码，20 FPS，256x256 |
observation.images.wrist |
video | (256, 256, 3) | 腕部视角 RGB 视频，AV1 编码，20 FPS，256x256 |
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 |
数据存储格式
- 数据路径:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频路径:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4
引用信息
引用信息暂缺。
