push-final-red-block-100-cutie-combined-master

该数据集由LeRobot创建，专注于机器人技术领域，具体针对so_follower型机器人。数据集包含100个剧集，总计21726帧，数据文件大小为100MB，视频文件大小为200MB，帧率为30fps。数据以parquet格式存储，视频以mp4格式存储。数据集特征包括动作（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、夹爪位置）、观测状态（与动作相同的关节位置）、来自两个摄像头视角（eagle和side）的图像观测（分辨率为240x320，3通道）、时间戳、帧索引、剧集索引等。数据集用于训练和评估机器人控制任务，支持机器人模仿学习或强化学习应用。

This dataset was developed by LeRobot, focusing on the robotics domain, specifically targeting so_follower-type robots. It contains 100 episodes, totaling 21726 frames, with structured data files sized at 100 MB and video files sized at 200 MB, operating at a frame rate of 30 fps. The structured data is stored in Parquet format, while the videos are encoded in MP4 format. The dataset includes the following features: actions (6 joint positions: shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist rotation, gripper position), observation states with the same joint positions as the actions, image observations from two camera viewpoints (eagle and side) with a resolution of 240×320 and 3 color channels, timestamps, frame indices, episode indices, and other relevant metadata. This dataset is used for training and evaluating robotic control tasks, supporting applications such as robotic imitation learning and reinforcement learning.