seriintan/eval_act_baseline_v4

该数据集是使用LeRobot创建的机器人学数据集，专门用于机器人控制与视觉任务。数据集包含12个完整的情节（episodes），总计2420帧数据，存储为parquet格式，并附带视频文件。数据特征包括：动作空间（6个关节位置：肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置）、观测状态（相同6个关节位置）、来自两个摄像头的图像观测（正面和侧面，分辨率均为480x640，3通道RGB，30fps），以及时间戳、帧索引、情节索引等元数据。数据集适用于机器人模仿学习、强化学习或视觉-动作映射等研究。

This dataset is a robotics dataset created using LeRobot, designed for robot control and vision tasks. It contains 12 complete episodes, totaling 2420 frames, stored in parquet format with accompanying video files. The features include: action space (6 joint positions: shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, and gripper position), observation state (same 6 joint positions), image observations from two cameras (front and side, both with 480x640 resolution, 3-channel RGB, 30fps), and metadata such as timestamp, frame index, episode index, etc. The dataset is suitable for research in robot imitation learning, reinforcement learning, or vision-action mapping.