congnguyenzn/soarm-101_20260527_192701
收藏资源简介:
该数据集是一个用于机器人学习的多模态数据集,由LeRobot工具创建。它包含一个总时长为1084帧的单个训练集,数据以30帧每秒的速率采集。数据集结构丰富,包括6维浮点型动作数据(如肩部、肘部、腕部和夹爪的位置控制)、6维浮点型观测状态数据(反映机器人关节位置),以及来自手腕和前置摄像头的视频观测(分辨率为480x640,3通道彩色视频,使用AV1编码)。此外,还提供了时间戳、帧索引、片段索引和任务索引等元数据,适用于机器人控制、模仿学习或强化学习任务。数据以Parquet文件格式存储,总数据大小约为300MB(其中视频数据占200MB)。机器人类型为so_follower,表明可能用于跟随或协作任务。
This dataset is a multimodal dataset for robotics learning, created using the LeRobot tool. It contains a single training set with a total of 1084 frames, collected at a rate of 30 frames per second. The dataset structure is comprehensive, including 6-dimensional float32 action data (e.g., position control for shoulder, elbow, wrist, and gripper), 6-dimensional float32 observation state data (reflecting robot joint positions), and video observations from wrist and front cameras (resolution 480x640, 3-channel color video, encoded with AV1). Additionally, it provides metadata such as timestamps, frame indices, episode indices, and task indices, making it suitable for robotics control, imitation learning, or reinforcement learning tasks. The data is stored in Parquet file format, with a total size of approximately 300MB (including 200MB for video data). The robot type is so_follower, indicating potential use in following or collaborative tasks.




