ccwatson/metaworld_spatial_diagonal_15cm_strict_50deg_close_high_approach_improved_mw

该数据集由LeRobot创建，包含机器人学习任务相关的数据。数据集结构包括多种特征，如观察状态、环境状态、图像（角落视角和夹持器视角）、动作、时间戳、帧索引、剧集索引等。具体来说，观察状态为4维浮点数，环境状态为39维浮点数，图像数据为224x224x3的尺寸。动作数据为4维浮点数。数据集共包含800个剧集，99144帧，8个任务，帧率为24fps。数据以parquet格式存储，分为1个块，每块1000个剧集。

This dataset was created using LeRobot and contains data related to robotics learning tasks. The dataset structure includes various features such as observation state, environment state, images (corner view and gripper view), actions, timestamp, frame index, episode index, etc. Specifically, the observation state is a 4-dimensional float, the environment state is a 39-dimensional float, and the image data is of size 224x224x3. The action data is a 4-dimensional float. The dataset consists of 800 episodes, 99144 frames, 8 tasks, with a frame rate of 24fps. The data is stored in parquet format, divided into 1 chunk with 1000 episodes per chunk.