taccap-g1-grap-fabric-0708
收藏Hugging Face2026-07-08 更新2026-07-08 收录
下载链接:
https://huggingface.co/datasets/TacVerse/taccap-g1-grap-fabric-0708
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是一个机器人数据集,使用LeRobot创建,专门用于双爪触觉抓取任务。它包含15个完整情节,总计79,972帧数据,采样率为30帧每秒。数据集采用Apache 2.0许可证。数据集中包括动作和观察状态特征,动作特征为20维浮点数组,表示左右机械臂的TCP(工具中心点)位置、旋转和夹爪位置;观察状态特征同样为20维浮点数组,提供机器人的实时状态信息。此外,数据集还包含多视角视觉数据:左右触觉传感器图像(分辨率400x700,3通道)和左右腕部摄像头图像(分辨率480x640,3通道),均以视频格式存储(H.264编码,YUV420p像素格式)。其他特征包括时间戳、帧索引、情节索引和任务索引。数据集结构以parquet文件组织,支持高效存储和访问,适用于机器人学习、触觉感知和强化学习研究。
This dataset is a robotics dataset created using LeRobot, specifically designed for dual-arm tactile grasping tasks. It contains 15 complete episodes, totaling 79,972 frames with a sampling rate of 30 frames per second. The dataset is licensed under Apache 2.0. It includes action and observation state features: action features are a 20-dimensional floating-point array representing the TCP (Tool Center Point) positions, rotations, and gripper positions of the left and right robotic arms; observation state features are also a 20-dimensional floating-point array, providing real-time state information of the robot. Additionally, the dataset contains multi-view visual data: left and right tactile sensor images (resolution 400x700, 3 channels) and left and right wrist camera images (resolution 480x640, 3 channels), all stored in video format (H.264 encoding, YUV420p pixel format). Other features include timestamps, frame indices, episode indices, and task indices. The dataset is organized in parquet files for efficient storage and access, suitable for research in robotics learning, tactile perception, and reinforcement learning.
提供机构:
TacVerse创建时间:
2026-07-08
原始信息汇总
数据集概述
- 数据集名称: taccap-g1-grap-fabric-0708
- 许可证: Apache-2.0
- 任务类别: 机器人学(Robotics)
- 标签: LeRobot
数据集结构
- 机器人类型: bi_taccap_gripper
- 总片段数: 15
- 总帧数: 79,972
- 总任务数: 1
- 帧率: 30 FPS
- 数据集分割: 仅包含训练集(0:15)
- 数据文件格式: Parquet
- 视频文件格式: MP4(H.264编码)
- 数据大小: 数据文件约100MB,视频文件约200MB
特征说明
数据集每条记录包含以下特征,均为浮点型(float32)或整型(int64),以及视频数据:
| 特征名称 | 数据类型 | 描述 |
|---|---|---|
action |
float32, 形状 [20] | 动作指令,包含左右TCP位置、姿态和夹爪位置 |
observation.state |
float32, 形状 [20] | 观察状态,与动作结构一致 |
observation.images.left_tactile_left |
video, 400x700x3 | 左侧触觉左相机图像 |
observation.images.left_tactile_right |
video, 400x700x3 | 左侧触觉右相机图像 |
observation.images.left_wrist |
video, 480x640x3 | 左侧手腕相机图像 |
observation.images.right_tactile_left |
video, 400x700x3 | 右侧触觉左相机图像 |
observation.images.right_tactile_right |
video, 400x700x3 | 右侧触觉右相机图像 |
observation.images.right_wrist |
video, 480x640x3 | 右侧手腕相机图像 |
timestamp |
float32, 形状 [1] | 时间戳 |
frame_index |
int64, 形状 [1] | 帧索引 |
episode_index |
int64, 形状 [1] | 片段索引 |
index |
int64, 形状 [1] | 全局索引 |
task_index |
int64, 形状 [1] | 任务索引 |
引用
如需引用该数据集,请使用以下BibTeX格式:
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} }



