robot-learning-group47/eval2_all_prompts

该数据集是使用LeRobot创建的，主要用于机器人学习任务。数据集包含267个episodes，总计160,200帧，帧率为30 fps。数据特征包括动作（6个浮点数值，对应机器人关节位置）、观测状态（6个浮点数值，同样对应关节位置）、观测图像（来自前摄像头的480x640 RGB图像，非深度图），以及时间戳、帧索引、episode索引等元数据。数据集分为15个任务，数据以parquet文件格式存储，总数据文件大小为100 MB，视频文件大小为200 MB。数据用于训练和评估机器人控制模型，支持强化学习或模仿学习应用。This dataset was created using LeRobot and is designed for robotics learning tasks. It consists of 267 episodes with a total of 160,200 frames at 30 fps. The features include actions (6 float32 values representing robot joint positions), observation states (6 float32 values for joint positions), observation images (front camera images of size 480x640 in RGB, not depth maps), along with metadata such as timestamp, frame index, episode index, etc. The dataset is divided into 15 tasks, stored in parquet format, with a total data file size of 100 MB and video file size of 200 MB. It is intended for training and evaluating robot control models, supporting applications in reinforcement learning or imitation learning.