L2D-v3

这是一个由LeRobot创建的数据集，是yaak-ai/L2D的R2版本（10K episodes），采用LeRobot Dataset V3格式。数据集包含车辆状态、车道、道路、表面、最大速度、降水、条件、照明、航点等信息，以及多个视角的图像观测。数据集的结构和特征在meta/info.json中有详细描述，包括观测状态、动作、时间戳、帧索引、任务索引等。This is a dataset created by LeRobot, which is the R2 version (10K episodes) of yaak-ai/L2D, adopting the LeRobot Dataset V3 format. The dataset contains information such as vehicle states, lanes, roads, surface conditions, maximum speed, precipitation, environmental conditions, lighting, waypoints, as well as image observations from multiple perspectives. The structure and characteristics of the dataset are detailed in meta/info.json, including observation states, actions, timestamps, frame indices, task indices and other relevant contents.

搜集汇总

背景与挑战

背景概述

L2D-v3是由LeRobot创建的数据集，为yaak-ai/L2D的R2版本，包含10K个episodes，采用LeRobot Dataset V3格式。该数据集涵盖了车辆状态、道路信息、环境条件及多视角图像观测等丰富特征，详细结构定义在meta/info.json中。