DriveLMM-o1
收藏资源简介:
DriveLMM-o1是一个专为自动驾驶场景理解设计的步进推理数据集,由 Mohamed Bin Zayed University of Artificial Intelligence 等机构创建。该数据集包含超过18000个训练集问题和4000多个测试集问题,覆盖了感知、预测和规划等多个方面的驾驶相关问题,每个问题都配备了逐步推理的注释。数据集整合了多视角图像和LiDAR点云,以支持模型对复杂驾驶场景的全面理解。
DriveLMM-o1 is a step-by-step reasoning dataset specifically designed for autonomous driving scenario understanding, developed by Mohamed Bin Zayed University of Artificial Intelligence and other institutions. This dataset includes over 18,000 training questions and more than 4,000 test questions, covering a variety of driving-related problems across multiple aspects such as perception, prediction and planning, and each question is provided with step-by-step reasoning annotations. The dataset integrates multi-view images and LiDAR point clouds to enable models to comprehensively understand complex driving scenarios.

- 1DriveLMM-o1: A Step-by-Step Reasoning Dataset and Large Multimodal Model for Driving Scenario Understanding Mohamed Bin Zayed University of Artificial Intelligence, Linköping University, Australian National University · 2025年



