HarikrishnaVydana/pick-place-redcube-3-cmaeras-complex-2-train
收藏资源简介:
该数据集是一个机器人学数据集,使用LeRobot创建,专注于机器人的拾取和放置任务(具体为pick-place-redcube-3-cameras-complex-2-train场景)。数据集包含丰富的多模态数据:动作数据为6维浮点向量,表示机器人关节位置(包括肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置);观察数据包括相同的6维状态向量和三个摄像头视角的视频(侧视、上视和腕部视角),每个视频分辨率为480x640,3通道,帧率为30fps,使用AV1编解码器。此外,数据集还包含时间戳、帧索引、episode索引、任务索引等元数据。数据集总共有126个episodes,93863帧,数据以parquet文件格式存储,视频以mp4格式存储,适用于机器人控制、强化学习或计算机视觉任务的研究和开发。
This dataset is a robotics dataset created using LeRobot, focusing on pick-and-place tasks (specifically the pick-place-redcube-3-cameras-complex-2-train scenario). It contains rich multimodal data: action data as a 6-dimensional float vector representing robot joint positions (including shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, and gripper positions); observation data includes the same 6-dimensional state vector and video from three camera perspectives (side, up, and wrist views), each with a resolution of 480x640, 3 channels, 30fps frame rate, using the AV1 codec. Additionally, the dataset includes metadata such as timestamps, frame indices, episode indices, and task indices. It comprises 126 episodes and 93,863 frames, stored in parquet format for data and mp4 format for videos, suitable for research and development in robot control, reinforcement learning, or computer vision tasks.




