taccap-g1-pick-up-QQ-candy-0707
收藏Hugging Face2026-07-07 更新2026-07-07 收录
下载链接:
https://huggingface.co/datasets/TacVerse/taccap-g1-pick-up-QQ-candy-0707
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操作数据集,专为双TacCap夹爪机器人(bi_taccap_gripper)设计。数据集包含总共61个episodes和140198帧数据,专注于单一任务,即拾取QQ糖果的操作。数据特征包括动作(action)和观察(observation),其中动作涉及左右夹爪的末端执行器位置(x, y, z)、姿态(r1到r6)以及夹爪位置(pos);观察包括机器人状态(与动作类似)和图像数据,图像来自左右手腕摄像头(分辨率480x640)和左右触觉传感器(分辨率400x700),提供视频流以捕捉视觉和触觉信息。数据集以30fps的帧率录制,数据存储为parquet文件和视频文件(mp4格式),适用于机器人学习、强化学习或触觉感知研究。数据集由LeRobot工具创建,并遵循Apache 2.0许可证。
This dataset is a robotic manipulation dataset designed for a bi_taccap_gripper robot. It contains a total of 61 episodes and 140,198 frames, focusing on a single task of picking up QQ candies. The data features include actions and observations: actions involve end-effector positions (x, y, z), orientations (r1 to r6), and gripper positions (pos) for both left and right grippers; observations include robot state (similar to actions) and image data from left and right wrist cameras (resolution 480x640) and left and right tactile sensors (resolution 400x700), providing video streams to capture visual and tactile information. The dataset is recorded at 30fps, stored as parquet files and video files (mp4 format), and is suitable for robot learning, reinforcement learning, or tactile perception research. It is created using the LeRobot tool and follows the Apache 2.0 license.
提供机构:
TacVerse创建时间:
2026-07-07



