HuskyMango/eval_diffusion-5-22

该数据集使用LeRobot创建，是一个用于机器人学习的单臂Franka机器人数据集。它包含动作数据（如位置、姿态和夹爪控制）、观测状态（如关节角度和夹爪状态）以及来自多个摄像头（包括手腕和场景摄像头）的视觉观测图像。数据以Parquet文件格式存储，帧率为20fps，图像分辨率为224x224。数据集结构包括时间戳、帧索引、episode索引和任务索引等特征，适用于机器人控制、视觉导航和强化学习等任务。

This dataset was created using LeRobot and is a single-arm Franka robot dataset for robotics learning. It includes action data (e.g., position, orientation, and gripper control), observation states (e.g., joint angles and gripper state), and visual observation images from multiple cameras (including wrist and scene cameras). The data is stored in Parquet format with a frame rate of 20fps and image resolution of 224x224. The dataset structure features timestamps, frame indices, episode indices, and task indices, making it suitable for tasks such as robot control, visual navigation, and reinforcement learning.