taccap-g1-organize-meeting-table-items-0729
收藏资源简介：
该数据集是一个专门用于双TacCap夹爪机器人执行组织会议桌面物品任务的机器人操作数据集，由LeRobot工具创建。它包含101个完整的情节，总计402,584帧数据，所有数据均用于训练。数据以parquet文件格式存储，总数据文件大小为100MB，视频文件大小为200MB，视频帧率为30fps。数据集的特征包括：20维的动作向量和状态观察，表示左右夹爪的工具中心点（TCP）的x、y、z位置、r1到r6的旋转参数以及夹爪位置；以及来自左右触觉传感器（左右各两个触觉图像，分辨率400x700）和左右手腕摄像头（分辨率480x640）的视频观察，所有视频均为彩色（3通道），使用h264编解码器。此外，还包含时间戳、帧索引、情节索引、全局索引和任务索引等元数据。该数据集适用于机器人学习、模仿学习或强化学习研究，特别适合涉及触觉感知和双爪操作的任务。
This dataset is a robotic manipulation dataset specifically designed for a dual TacCap gripper robot performing the organizing meeting tabletop objects task. It was created using the LeRobot tool and contains 101 complete episodes, totaling 402,584 frames, all of which are used for training. The data is stored in parquet format, with a total data file size of 100MB and a video file size of 200MB, with a video frame rate of 30fps. The dataset features include: a 20-dimensional action vector and state observation, representing the x, y, z positions, r1 to r6 rotation parameters, and gripper positions of the left and right gripper tool center points (TCP); as well as video observations from left and right tactile sensors (two tactile images each, with a resolution of 400x700) and left and right wrist cameras (resolution 480x640), all videos are in color (3 channels) and use the h264 codec. Additionally, it includes metadata such as timestamps, frame indices, episode indices, global indices, and task indices. The dataset is suitable for robotics learning, imitation learning, or reinforcement learning research, particularly for tasks involving tactile perception and dual-gripper manipulation.
数据集概述：taccap-g1-organize-meeting-table-items-0729
- 数据集地址：https://huggingface.co/datasets/TacVerse/taccap-g1-organize-meeting-table-items-0729
- 许可证：Apache-2.0
- 任务类别：机器人学（robotics）
- 标签：LeRobot
- 创建工具：LeRobot (https://github.com/huggingface/lerobot)
数据集结构
基础信息
| 属性 | 值 |
|---|---|
| 代码库版本 | v3.0 |
| 机器人类型 | bi_taccap_gripper |
| 总片段数 | 101 |
| 总帧数 | 402,584 |
| 总任务数 | 1 |
| 数据文件大小 | 100 MB |
| 视频文件大小 | 200 MB |
| FPS | 30 |
| 数据分片 | 训练集：片段 0 至 100 |
数据文件格式
- 数据文件：Parquet 格式，路径为
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频文件：MP4 格式（H.264 编码），路径为
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4
特征说明
| 特征名称 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
| action | float32 | (20,) | 包含左右 TCP 位置、旋转及夹爪位置的动作指令 |
| observation.state | float32 | (20,) | 与 action 对应的机器人状态观测 |
| observation.images.left_tactile_left | video | (400, 700, 3) | 左侧触觉左侧摄像头视频，FPS 30，H.264 编码 |
| observation.images.left_tactile_right | video | (400, 700, 3) | 左侧触觉右侧摄像头视频 |
| observation.images.left_wrist | video | (480, 640, 3) | 左侧手腕摄像头视频 |
| observation.images.right_tactile_left | video | (400, 700, 3) | 右侧触觉左侧摄像头视频 |
| observation.images.right_tactile_right | video | (400, 700, 3) | 右侧触觉右侧摄像头视频 |
| observation.images.right_wrist | video | (480, 640, 3) | 右侧手腕摄像头视频 |
| 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 |
引用
bibtex @misc{xense-taccap-lerobot, author = {XenseRobotics Team}, title = {LeRobot-Xense: LeRobot with Xense Tactile Robotics Support}, howpublished = {url{https://github.com/Vertax42/xense-taccap-lerobot}}, year = {2026} }
