droid_place_hammer

该数据集是一个机器人数据集，使用LeRobot工具创建，专门用于机器人学习任务。数据集基于Franka机器人，包含20个完整的情节（episodes），总计1568帧数据，针对单一任务设计。数据以parquet文件格式存储，并附带视频文件，帧率为10 fps。数据集特征丰富，包括观察状态（如关节位置、关节速度、笛卡尔位置、夹爪位置）、动作、奖励、完成标志和成功标志，以及来自左腕摄像头的图像数据（分辨率为720x1280，3通道）。数据集适用于机器人控制、强化学习等研究领域，遵循Apache 2.0许可证。

This dataset is a robotics dataset created using the LeRobot tool, specifically designed for robot learning tasks. It is based on the Franka robot, containing 20 complete episodes with a total of 1568 frames of data, tailored for a single task. The data is stored in parquet file format and includes video files with a frame rate of 10 fps. The dataset features rich information, including observation states (such as joint positions, joint velocities, Cartesian positions, gripper positions), actions, rewards, done flags, success flags, and image data from the left wrist camera (resolution 720x1280, 3 channels). It is suitable for research areas such as robot control and reinforcement learning, and follows the Apache 2.0 license.