nsimonato25/libero_no_ketchup

该数据集使用LeRobot创建，专为机器人技术任务设计。数据集包含1296个episodes，总计169265帧，覆盖29个不同任务。数据以10fps的帧率采集，包括来自两个摄像头（image和image2）的256x256 RGB图像观测、8维状态观测和7维动作数据。数据集结构分为训练集，数据文件以parquet格式存储，总大小约100MB，视频文件总大小约200MB。数据集适用于机器人控制、强化学习等应用，支持基于图像和状态输入的机器人行为学习。This dataset was created using LeRobot and is designed for robotics tasks. It contains 1296 episodes with a total of 169265 frames, covering 29 different tasks. Data is collected at 10fps and includes observations from two cameras (image and image2) with 256x256 RGB images, 8-dimensional state observations, and 7-dimensional action data. The dataset structure includes a training split, with data files stored in parquet format totaling approximately 100MB, and video files totaling approximately 200MB. It is suitable for applications such as robot control and reinforcement learning, supporting robot behavior learning based on image and state inputs.