datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
nuscenes-qa-mini
NuScenes-QA-mini Dataset
TL;DR:
This dataset is used for multimodal question-answering tasks in autonomous driving scenarios. We created this dataset based on nuScenes-QA dataset for evaluation in our paper Modality Plug-and-Play: Elastic Modality Adaptation in Multimodal LLMs for Embodied AI. The samples are divided into day and night scenes.
scene
# train samples
# validation samples
day
2,229
2,229
night
659
659
Each sample contains… See the full description on the dataset page: https://huggingface.co/datasets/KevinNotSmile/nuscenes-qa-mini.Ours-V2_nuscenes_ArmGS
Ours-V2_nuscenes_ArmGS
Preprocessing artifacts for the 10 selected scenes/logs used by this ArmGS workspace. This is not the entire upstream dataset and not an evaluation/trained-model repository.
Includes selected-scene sky masks, COLMAP mappings and final triangulated point clouds. The original nuScenes v1.0-trainval RGB/LiDAR/metadata are not included and remain required by the matching ArmGS raw-data adapter.
Layout and use
Files retain their… See the full description on the dataset page: https://huggingface.co/datasets/Dororo99/Ours-V2_nuscenes_ArmGS.Single-DriveLM-NuScenes-VQA
Single-DriveLM-NuScenes VQA Dataset
Updates & News
[03/05/2025] Our latest Trustworthy VLM benchmark AUTOTRUST was build on this dataset
[10/11/2024] VQA Dataset was released
Dataset Description
This is the sub-dataset of DriveLM which only include single object in ego scenes
Uses
For single traffic participant recgonition, segmentation, VQA subtasks of driving scenarios.
Dataset Structure
single_pedestrian
├── images
└──… See the full description on the dataset page: https://huggingface.co/datasets/Chouoftears/Single-DriveLM-NuScenes-VQA.nuscenes_qa_groupedMoRAL-nuscenes-v3trajectory-prediction-nuscenesNuscenes_AutoMoTMoRAL-nuscenesNuscenes_depth_estimationNuscenes-v1.0-trainval-CAM_FRONTnuScenes-Geography-Data
Spatial Retrieval Augmented Autonomous Driving
For more details about the dataset and the project, please visit Spatial Retrieval Augmented Autonomous Driving.
nuScenes-Geography Dataset
nuScenes-Geography
├── frame_metadata.json
├── pano_metadata.json
├── unavailable_metadata.json
├── sat
│ ├── boston-seaport.png
│ ├── singapore-hollandvillage.png
│ ├── singapore-onenorth.png
│ └── singapore-queenstown.png
└── streetview
├── quality_labels.json
└── panos… See the full description on the dataset page: https://huggingface.co/datasets/SpatialRetrievalAD/nuScenes-Geography-Data.nuScenes_raw_datanuscenes_frontNuscenesQA_raw_datacarla-nuscenesMagicDriveDiT-nuScenes-metadata
MagicDriveDiT
📄 Paper |
🌐 Website |
📖 LICENSE |
🤖 GitHub
This repository contains the pre-processed metadata for nuScenes used in the paper.
MagicDriveDiT: High-Resolution Long Video Generation for Autonomous Driving with Adaptive Control
Ruiyuan Gao1, Kai Chen2, Bo Xiao3, Lanqing Hong4, Zhenguo Li4, Qiang Xu1
1CUHK 2HKUST 3Huawei Cloud 4Huawei Noah's Ark Lab
Please find more information on our GitHub:… See the full description on the dataset page: https://huggingface.co/datasets/flymin/MagicDriveDiT-nuScenes-metadata.Nuscenes-QA-merge-front-imageUSAGE in Python
load train and valid dataset
add base_folder
Nuscenesocc3d-nuscenes-transferMoRAL-nuscenes-v2moral-v4-nuscenesnuScenes-UNION-labels
UNION: Unsupervised 3D Object Detection using Appearance-based Pseudo-Classes [NeurIPS 2024]
[arXiv] [GitHub] [BibTeX]
Hugging Face repository for prebuilt .pkl annotation files created by UNION on the nuScenes dataset.
This repo is annotations-only (no sensor data) so you can reproduce UNION results quickly without regenerating labels.
🔗 Code: https://github.com/TedLentsch/UNION
⬇️ nuScenes download: https://www.nuscenes.org/nuscenes
🛡️ License: CC BY-NC-SA 4.0 (inherits… See the full description on the dataset page: https://huggingface.co/datasets/TedLentsch/nuScenes-UNION-labels.Gen-nuScenesOpenDataLab___nuScenes
数据集介绍
简介
nuScenes数据集是一个大规模的自动驾驶数据集。该数据集具有用于在波士顿和新加坡收集的1000场景的3D边界框。每个场景长20秒，注释为2Hz。这导致总共28130个用于训练的样本，6019个用于验证的样本和6008个用于测试的样本。该数据集具有完整的自动驾驶车辆数据套件: 32光束激光雷达，6个摄像头和具有完整360 ° 覆盖的雷达。3D对象检测挑战评估10个类别的性能: 汽车，卡车，公共汽车，拖车，建筑车辆，行人，摩托车，自行车，交通锥和障碍物。
类定义
animal
human.pedestrian.adult
human.pedestrian.child
human.pedestrian.construction_worker
human.pedestrian.personal_mobility
human.pedestrian.police_officer
human.pedestrian.stroller
human.pedestrian.wheelchair… See the full description on the dataset page: https://huggingface.co/datasets/AlayaNeW/OpenDataLab___nuScenes.nuscenes-metanuscenesqaNuscenes-v1.0-trainval-CAM_FRONT_SweepsNuscenes-QA-all-imagesNuScenes-Asianuscenes-compressed