HSJUSER/ffw_sg2_rev1_0526_diffusion_Test_a1

该数据集是一个机器人学数据集，使用LeRobot创建，专门用于机器人控制和模仿学习任务。数据集包含9个episodes，总计5405帧和27个视频，覆盖单一任务。数据以30帧每秒的速率采集，并分为训练集（全部9个episodes）。数据集的结构包括多个特征：时间戳、帧索引、episode索引、任务索引，以及来自三个摄像头的图像观测（头部摄像头：分辨率376x672x3，左腕摄像头：240x424x3，右腕摄像头：240x424x3）。此外，还提供了机器人状态观测，包含19个关节位置（如左臂关节、右臂关节、头部关节和升降关节），以及相应的19维动作向量，用于控制这些关节。数据集以Apache 2.0许可证发布，适用于机器人行为克隆、强化学习等研究。This dataset is a robotics dataset created using LeRobot, specifically designed for robot control and imitation learning tasks. It contains 9 episodes, totaling 5405 frames and 27 videos, covering a single task. Data is collected at 30 frames per second and split into a training set (all 9 episodes). The dataset structure includes multiple features: timestamp, frame index, episode index, task index, as well as image observations from three cameras (head camera: resolution 376x672x3, left wrist camera: 240x424x3, right wrist camera: 240x424x3). Additionally, it provides robot state observations comprising 19 joint positions (e.g., left arm joints, right arm joints, head joints, and lift joint), along with corresponding 19-dimensional action vectors for controlling these joints. The dataset is released under the Apache 2.0 license and is suitable for research in robot behavior cloning, reinforcement learning, and related areas.