dmody1/collect5
收藏资源简介:
该数据集是一个机器人或自动驾驶相关数据集,包含多个特征字段:episode_id(剧集ID)、frame_idx(帧索引)、image(图像)、odom_0到odom_12(里程计数据,可能表示位置、方向或速度信息)、action_0到action_13(动作命令数据,可能控制机器人的运动)、imu_0到imu_9(惯性测量单元数据,如加速度和角速度)、depth_mean(平均深度值)和source(数据来源字符串)。数据集仅包含训练分割,共有633,050个示例,总大小约为20.7GB,适用于机器学习和计算机视觉任务,如轨迹预测、动作学习或环境感知。
This dataset is related to robotics or autonomous driving, featuring multiple fields: episode_id, frame_idx, image, odom_0 to odom_12 (odometry data, possibly representing position, orientation, or velocity information), action_0 to action_13 (action command data, likely controlling robot motion), imu_0 to imu_9 (inertial measurement unit data, such as acceleration and angular velocity), depth_mean (average depth value), and source (data source string). The dataset includes only a training split with 633,050 examples and a total size of approximately 20.7GB, suitable for machine learning and computer vision tasks like trajectory prediction, action learning, or environmental perception.



