datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Debunk_Traffic_Representation
Packet-level classification: classify based on packet
Per-packet-split: Mix all packets and split them into train, val, and test sets, based on 8:1:1
Per-flow-split: Split the pcap files based on 5-tuples (src_IP, dst_IP, src_port, dst_port, and protocol), using 3-fold validation, and there is no intersection between the train, val, and test sets.
Flow-level classification: classify based on flow
traffic-vehicle-detection
Edge-AI Traffic Vehicle Detection (UA-DETRAC CCTV)
Part of the Edge-AI Traffic & Vehicle Analytics System repository by thundarstrom.
Dataset Summary
Curated and normalized 23,319 CCTV traffic images from fixed intersection surveillance cameras (UA-DETRAC benchmark). Contains 215,109 annotated bounding boxes in standard YOLO format across 4 vehicle classes: car, bus, truck, and van.
Class Mapping
Class 0 (car): 177,403 bboxes (82.5%)
Class 1 (bus):… See the full description on the dataset page: https://huggingface.co/datasets/PRAS4NTH/traffic-vehicle-detection.vehicular-traffic-light-reinforcementtrafficmanagementdataset
Traffic Management Dataset
A large-scale urban traffic dataset combining Bengaluru CCTV footage with the UA-DETRAC benchmark. Contains annotated images for vehicle detection, classification, and multi-object tracking across four sub-datasets totalling ~196,000 images.
Dataset Structure
├── BMD-45-Train/ # 35,792 images, COCO format annotations
│ ├── images_000/
│ └── _annotations.coco.json
├── UVH-26-Train/ # 17,387 images, COCO format… See the full description on the dataset page: https://huggingface.co/datasets/kalyan1729/trafficmanagementdataset.TrafficQAbrb-traffic-videos_ORIGINAL1brb-traffic-videos_ORIGINAL2brb-traffic-videos_ORIGINAL4brb-traffic-videos_ORIGINAL3traffic-detectionLISA-Traffic-Lights
LISA Traffic Light Dataset (YOLO Box Export)
Unofficial redistribution of the LISA Traffic Light Dataset's "box" annotation style, exported in YOLO format, under the original CC BY-NC-SA 4.0 license.
Disclaimer
This repository is not an official release of the LISA Traffic Light Dataset.
The dataset was created by Morten Bornø Jensen, Mark Philip Philipsen, Andreas Møgelmose, Thomas B. Moeslund, and Mohan M. Trivedi at the Laboratory for Intelligent… See the full description on the dataset page: https://huggingface.co/datasets/dronefreak/LISA-Traffic-Lights.brb-traffic-videostraffic_signal_imagesThis dataset contains traffic images from traffic signal cameras of singapore. The images are captured at 1.5 minute interval from 6 pm to 7 pm everyday for the month of January 2024.road-traffic
Road Traffic
This dataset is part of the Roboflow 100 benchmark, a diverse collection of 100 object detection datasets spanning 7 imagery domains.
Dataset Statistics
Split
Images
Train
494
Validation
187
Test
133
Total
814
Classes (12)
bicycles
bridges
buses
chimneys
crosswalks
fire hydrants
motorcycles
stairs
taxis
tractors
traffic lights
vehicles
Usage
With LibreYOLO
from libreyolo import LIBREYOLO
# Load a model… See the full description on the dataset page: https://huggingface.co/datasets/LibreYOLO/road-traffic.vehicle-mixed-traffic-detection
Visaitech Mixed-Traffic Vehicle Detection Dataset (v0.1)
Dashcam frames annotated for pedestrian / 2-wheeler / 3-wheeler / 4-wheeler
detection in South Asian mixed traffic, a class taxonomy general-purpose
COCO-trained detectors don't cover (COCO has no concept of an auto-rickshaw
or motorcycle-vs-bicycle-as-one-class "2-wheeler" grouping tuned for how
this traffic actually mixes on the road).
This is an early v0.1 release: 293 annotated frames from 6 source videos,
published… See the full description on the dataset page: https://huggingface.co/datasets/visaitech/vehicle-mixed-traffic-detection.traffickaggle_web_traffic
kaggle_web_traffic (TsFile format)
145063 daily time series representing the number of hits or web traffic for a set of Wikipedia pages from 2015-07-01 to 2017-09-10.
This repository contains the full source .tsf series from the Monash Time Series Forecasting Repository converted to Apache TsFile format.
Summary
Source dataset: Monash-University/monash_tsf
Original source: https://zenodo.org/record/4656080
Monash subset: kaggle_web_traffic
Modalities: Time-series… See the full description on the dataset page: https://huggingface.co/datasets/THULab/kaggle_web_traffic.traffic-sign-bench
Traffic Sign Bench
Official per-sign SUMO maps for TrafficRuleBench: real Moscow OSM
layouts, 25 signs, 2500 maps. Protocol size is
80 train + 20 test maps per sign.
Road geometry is derived from OpenStreetMap
© OpenStreetMap contributors and is released under ODbL 1.0.
Download
All scenes land under data/scenes/<sign>/<scene_id>/, which is what eval
expects:
huggingface-cli download emb-ai/traffic-sign-bench \
--repo-type dataset \… See the full description on the dataset page: https://huggingface.co/datasets/emb-ai/traffic-sign-bench.air-traffictraffic-vehicle-detection
Edge-AI Traffic Vehicle Detection (UA-DETRAC CCTV)
Part of the Edge-AI Traffic & Vehicle Analytics System repository by thundarstrom.
Dataset Summary
Curated and normalized 23,319 CCTV traffic images from fixed intersection surveillance cameras (UA-DETRAC benchmark). Contains 215,109 annotated bounding boxes in standard YOLO format across 4 vehicle classes: car, bus, truck, and van.
Class Mapping
Class 0 (car): 177,403 bboxes (82.5%)
Class 1 (bus):… See the full description on the dataset page: https://huggingface.co/datasets/thundarstrom/traffic-vehicle-detection.TrafficForecast-LargeST-Project-Dataset
数据目录结构
TrafficForecast-LargeST/├─ data/│ ├─ raw/ # 原始数据（HDF5、metadata、adjacency）│ ├─ interim/ # 预处理中间结果（可选）│ └─ processed/ # 已处理数据缓存（.npz）
PhysicalAI-Traffic-Anomaly-Reasoning
Traffic Anomaly Reasoning (TAR)
This dataset is the official release for AI City Challenge 2026
Track 3 — Anomalous Events in Transportation.
It contains:
Training (train/): 44,040 pseudo-labeled multi-task annotations
covering 3,670 CCTV videos (≈26.1 hours: 9.2 hrs anomalous + 16.9 hrs
normal) sourced from eight public datasets.
Test (test/): 960 human-curated annotations covering 80 short
clips trimmed from 17 public YouTube videos. Answers are redacted in
this release;… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/PhysicalAI-Traffic-Anomaly-Reasoning.NYC_traffic_speed
Data source:
NYC_Traffic_Speed
Dataset Structure
The dataset is organized into the following structure:
|-- subdataset1
| |-- raw_data # Original data files
| |-- time_series # Rule-based Imputed data files
| | |-- id_1.parquet # Time series data for each subject can be multivariate, can be in csv, parquet, etc.
| | |-- id_2.parquet
| | |-- ...
| | |-- id_info.json # Metadata for each subject
| |-- weather
| | |-- location_1
| | | |--… See the full description on the dataset page: https://huggingface.co/datasets/fidel-ts/NYC_traffic_speed.VN_TrafficBuddy_VQAmoroccan-traffic-sign-datasettraffic_hourly
traffic_hourly (TsFile format)
862 hourly time series showing the road occupancy rates on the San Francisco Bay area freeways from 2015 to 2016.
This repository contains the full source .tsf series from the Monash Time Series Forecasting Repository converted to Apache TsFile format.
Summary
Source dataset: Monash-University/monash_tsf
Original source: https://zenodo.org/record/4656132
Monash subset: traffic_hourly
Modalities: Time-series
Source series: 862
Rows:… See the full description on the dataset page: https://huggingface.co/datasets/THULab/traffic_hourly.grand-bassin-traffic
Grand Bassin Traffic Dataset
An anonymised, auto-annotated image dataset of vehicle and pedestrian traffic
during the Maha Shivaratri pilgrimage at Grand Bassin (Ganga Talao), Mauritius,
captured simultaneously from an overhead (near-nadir) parking-area camera and a
roadside CCTV camera. The paired viewpoints make it suitable for studying the
aerial-vs-ground domain gap in object detection; the companion dataset
(Grand Bassin Kanwar Dataset) covers the pilgrimage-specific… See the full description on the dataset page: https://huggingface.co/datasets/shivam11/grand-bassin-traffic.Piper_traffic_light_water_color_merged
Piper Merged LeRobot Dataset
Generated by data/utils/merge_piper_hdf5_lerobot_to_lerobot.py.
Repo id: ZhaoRunyi/Piper_traffic_light_water_color_merged
FPS: 10
Visual backend: video
LeRobot merge mode: copy
State/action order: per arm joint6, gripper01, ee_pos3, ee_rot6d6.
Input paths:
/home/edemlab/challenge_ws/data/ZhaoRunyi/Piper_traffic_light_water_color
/home/edemlab/challenge_ws/data/challenge_hdf5/PourBeaker_Long_color_0627
observation.state and action use identical… See the full description on the dataset page: https://huggingface.co/datasets/ZhaoRunyi/Piper_traffic_light_water_color_merged.german-traffic-sign-detection
Dataset Labels
['animals', 'construction', 'cycles crossing', 'danger', 'no entry', 'pedestrian crossing', 'school crossing', 'snow', 'stop', 'bend', 'bend left', 'bend right', 'give way', 'go left', 'go left or straight', 'go right', 'go right or straight', 'go straight', 'keep left', 'keep right', 'no overtaking', 'no overtaking -trucks-', 'no traffic both ways', 'no trucks', 'priority at next intersection', 'priority road', 'restriction ends', 'restriction ends -overtaking… See the full description on the dataset page: https://huggingface.co/datasets/keremberke/german-traffic-sign-detection.NYC_traffic_speed
WIATS: Weather-centric Intervention-Aware Time Series Multimodal Dataset
Data source:
NYC_Traffic_Speed
Dataset Structure
The dataset is organized into the following structure:
|-- subdataset1
| |-- raw_data # Original data files
| |-- time_series # Rule-based Imputed data files
| | |-- id_1.parquet # Time series data for each subject can be multivariate, can be in csv, parquet, etc.
| | |-- id_2.parquet
| | |-- ...
| | |-- id_info.json… See the full description on the dataset page: https://huggingface.co/datasets/VEWOXIC/NYC_traffic_speed.