CoRL2026-CSI/UR7e_CaP_MultiTask_300epi_10fps_state_tplus1_action

这是一个公开的LeRobot v3.0多任务数据集，专为UR7e机器人的Code-as-Policies操作设计。数据集合并了三个100-episode 10fps数据集，形成一个包含300个episodes的训练语料库，保留了原始的数字观察、动作和任务标签。上传的最终版本将所有视频存储为10fps H.264 MP4文件。数据集包括来自顶部视角和手腕视角摄像头的图像观察（分辨率480x640 RGB），动作数据为绝对7D UR7e关节/夹爪目标，以及任务指令和每帧技能文本。具体任务包括：将红、绿、蓝积木从左到右沿紫色线排列；在蓝色盘子上堆叠积木，底部为红色积木，中间为绿色积木，顶部为蓝色积木；拾取红色积木并放置到蓝色盘子上。数据集来源于三个Hub仓库，并进行了动作重新标记，其中动作列从未来观察到的机器人状态重写。This is a public LeRobot v3.0 multi-task dataset for UR7e Code-as-Policies manipulation. It merges three 100-episode 10fps datasets into one 300-episode training corpus while preserving the original numeric observations, actions, and task labels. The final uploaded copy stores all videos as 10fps H.264 MP4 files. The dataset includes image observations from top-view and wrist cameras (480x640 RGB), actions as absolute 7D UR7e joint/gripper targets, and task instructions with per-frame skill text. Specific tasks involve: arranging red, green, blue blocks along a purple line from left to right; building a stack on a blue dish with the red block on the bottom, green in the middle, and blue on top; and picking up a red block and placing it on a blue dish. The dataset is derived from three Hub repositories and includes action relabeling where the action column is rewritten from future observed robot state.