datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
TrackingNet
TrackingNet devkit
This repository contains the data of the paper TrackingNet: A Large-Scale Dataset and Benchmark for Object Tracking in the Wild.
Download from HuggingFace
Download splits
from huggingface_hub import snapshot_download
# Download TRAIN_0 split (90GB)
snapshot_download(repo_id="SilvioGiancola/TrackingNet",
repo_type="dataset", revision="main",
local_dir="TrackingNet_HF"… See the full description on the dataset page: https://huggingface.co/datasets/SilvioGiancola/TrackingNet.ship-tracking-datahand_tracking_challenge_umetrackCheck out the Multiview Egocentric Hand Tracking Challenge 2024!!
To use this dataset, check out the hand_tracking_toolkit
visem-tracking-graphs
VISEM-Tracking-graphs - HuggingFace Repository
This HuggingFace repository contains the pre-generated graphs for the sperm video dataset called VISEM-Tracking (https://huggingface.co/papers/2212.02842) . The graphs represent spatial and temporal relationships between sperm in a video. Spatial edges connect sperms within the same frame, while temporal edges connect sperms across different frames.
The graphs have been generated with varying spatial threshold values: 0.1, 0.2, 0.3, 0.4… See the full description on the dataset page: https://huggingface.co/datasets/SimulaMet-HOST/visem-tracking-graphs.Maritime_Visual_Tracking_Dataset_MVTD
MVTD: Maritime Visual Tracking Dataset
Overview
MVTD (Maritime Visual Tracking Dataset) is a large-scale benchmark dataset designed specifically for single-object visual tracking (VOT) in maritime environments.It addresses challenges unique to maritime scenes: such as water reflections, low-contrast objects, dynamic backgrounds, scale variation, and severe illumination changes—which are not adequately covered by generic tracking datasets.
The dataset contains 182… See the full description on the dataset page: https://huggingface.co/datasets/AhsanBB/Maritime_Visual_Tracking_Dataset_MVTD.Maritime_Visual_Tracking_Dataset_MVTD
MVTD: Maritime Visual Tracking Dataset
Overview
MVTD (Maritime Visual Tracking Dataset) is a large-scale benchmark dataset designed specifically for single-object visual tracking (VOT) in maritime environments.It addresses challenges unique to maritime scenes: such as water reflections, low-contrast objects, dynamic backgrounds, scale variation, and severe illumination changes—which are not adequately covered by generic tracking datasets.
The dataset contains 182… See the full description on the dataset page: https://huggingface.co/datasets/othmaneirl/Maritime_Visual_Tracking_Dataset_MVTD.human-motion-tracking-deeplabcutThis dataset is used to adapt DeepLabCut for Human motion tracking.
Structure of the dataset
videos contains 100+ videos of 4 candidates recorded during a game of darts.
labeled-data contains labels on the corresponding frames of the videos. These labels are used to adapt DeepLabCut for human motion tracking. Under labeled-data there are 2 folders for every video.
video_name has all the relevant frames extracted from the video, xy coordinates of the labels in the csv file and the… See the full description on the dataset page: https://huggingface.co/datasets/GT-Neuronext/human-motion-tracking-deeplabcut.human-motion-tracking-deeplabcutThis dataset is used to adapt DeepLabCut for Human motion tracking.
Structure of the dataset
videos contains 100+ videos of 4 candidates recorded during a game of darts.
labeled-data contains labels on the corresponding frames of the videos. These labels are used to adapt DeepLabCut for human motion tracking. Under labeled-data there are 2 folders for every video.
video_name has all the relevant frames extracted from the video, xy coordinates of the labels in the csv file and the… See the full description on the dataset page: https://huggingface.co/datasets/pratikshapai/human-motion-tracking-deeplabcut.MUOT_3M-A_3_Million_Frame_Underwater_Object_Tracking_Dataset
🌊 MUOT-3M: The Largest Multimodal Underwater Object Tracking Dataset
Official repository for MUOT-3M📄 MUOT-3M: The Largest Multimodal Underwater Object Tracking Dataset and MUTrack Tracking Method
🚀 Overview
MUOT-3M is currently the largest underwater object tracking dataset, containing over 3 million annotated frames across 3,030 underwater videos with synchronized multimodal annotations.
The benchmark is designed to advance research in:
Underwater object tracking… See the full description on the dataset page: https://huggingface.co/datasets/AhsanBB/MUOT_3M-A_3_Million_Frame_Underwater_Object_Tracking_Dataset.evaldt
Prepared Dataset Archives
This repository stores a prepared split as tar.zst archives to avoid
millions of small files in the Hugging Face repository.
Extraction:
mkdir -p data
for f in *.tar.zst; do
tar --use-compress-program=zstd -xf "$f" -C data
done
sha256sum -c SHA256SUMS
After extraction, the layout is:
data/
dataset.json
dataset_stats.json
jsonl/
frames/
vision_cache/
gimbaled-uav-tracking-dataset
Gimbaled UAV Tracking Dataset / 云台无人机追踪数据集
A real-world, multi-sensor dataset for active localization of a non-cooperative
UAV using a two-axis gimbaled LiDAR–camera fusion system. It contains 16 flight
sequences acquired with a ground vehicle platform, together with the associated
tracking outputs and an RTK-based position reference.
面向非合作无人机主动定位的实测多传感数据集，采集自一套基于两轴云台
LiDAR–相机融合的地面车辆平台。包含 16 个飞行序列，并附带相应的跟踪
输出与基于 RTK 的位置参考。
The sequences span two acquisition days, sunny and… See the full description on the dataset page: https://huggingface.co/datasets/humanoidro/gimbaled-uav-tracking-dataset.evalstt
Prepared Dataset Archives
This repository stores a prepared split as tar.zst archives to avoid
millions of small files in the Hugging Face repository.
Extraction:
mkdir -p data
for f in *.tar.zst; do
tar --use-compress-program=zstd -xf "$f" -C data
done
sha256sum -c SHA256SUMS
After extraction, the layout is:
data/
dataset.json
dataset_stats.json
jsonl/
frames/
vision_cache/
NJU-HARD-Tracking
NJU-HARD-Tracking
Multi-Object Tracking across 122 MP UAV Image Sequences
🤗 Hugging Face · 🟣 ModelScope · 📊 Statistics: HF / MS
English | 中文: Hugging Face · ModelScope
🌍 Overview
NJU-HARD-Tracking provides the multi-object-tracking release of HARD, with full-resolution frames, temporal ordering, and the original instance annotations. It supports studying how detection and association behave across wide-area aerial… See the full description on the dataset page: https://huggingface.co/datasets/RL-MIND/NJU-HARD-Tracking.SN-Tracking-2023pig-detection-and-tracking
🐷 Dataset for the paper: Benchmarking pig detection and tracking under diverse and challenging conditions
Note: This dataset does not provide a load_dataset interface.
More detailed descriptions about the dataset can be found in the corresponding paper.
⬇️ Downloading the dataset
You can download the dataset using the Hugging Face CLI. First, install the following package:
pip install huggingface_hub
The entire dataset (roughly 25 GB) can then be downloaded as follows:… See the full description on the dataset page: https://huggingface.co/datasets/jonaden/pig-detection-and-tracking.reid-people-tracking
Reid People Tracking
Person re-identification dataset for tracking individuals across camera views.
Details
Property
Value
Samples
101
Media
image
Fields
global_id, confidence, camera_id, timestamp, embeddings
Usage
from fiftyone.utils.huggingface import load_from_hub
dataset = load_from_hub("Voxel51/reid-people-tracking")
Uploaded with FiftyOne Skills
cars-object-tracking
Cars Object Tracking
Dataset comprises 10,000+ video frames featuring both light vehicles (cars) and heavy vehicles (minivans). This extensive collection is meticulously designed for research in multi-object tracking and object detection, providing a robust foundation for developing and evaluating various tracking algorithms for road safety system development.
By utilizing this dataset, researchers can significantly enhance their understanding of vehicle dynamics and improve… See the full description on the dataset page: https://huggingface.co/datasets/UniDataPro/cars-object-tracking.VISEM-Tracking
Dataset Card for VISEM-Tracking
Dataset Summary
VISEM-Tracking is a dataset of human spermatozoa tracking, featuring 20 video recordings of 30 seconds each, captured from wet semen preparations. The dataset contains 29,196 frames with manually annotated bounding-box coordinates and a set of sperm characteristics analyzed by domain experts. In addition to the labeled dataset, unlabeled videos are provided to enable research in self-supervised learning. This dataset serves… See the full description on the dataset page: https://huggingface.co/datasets/sperm-net/VISEM-Tracking.pitch-control-tracking
Pitch Control Tracking Data
Per-player per-frame pitch control values from ~38 million rows of professional soccer tracking data across 20 matches from three providers. Computed using the Spearman (2017) physics-based model — each row contains one player's position, velocity, and the home-team control probability at that location.
Part of the (Right! Luxury!) Lakehouse soccer analytics platform.
Quick Start
from datasets import load_dataset
ds =… See the full description on the dataset page: https://huggingface.co/datasets/luxury-lakehouse/pitch-control-tracking.yuxuan_good_dataset_dtVISEM-Tracking
Dataset Card for VISEM-Tracking
Dataset Summary
VISEM-Tracking is a dataset of human spermatozoa tracking, featuring 20 video recordings of 30 seconds each, captured from wet semen preparations. The dataset contains 29,196 frames with manually annotated bounding-box coordinates and a set of sperm characteristics analyzed by domain experts. In addition to the labeled dataset, unlabeled videos are provided to enable research in self-supervised learning. This dataset… See the full description on the dataset page: https://huggingface.co/datasets/LandyJJ/VISEM-Tracking.Maritime_Visual_Tracking_Dataset_MVTD
MVTD: Maritime Visual Tracking Dataset
Overview
MVTD (Maritime Visual Tracking Dataset) is a large-scale benchmark dataset designed specifically for single-object visual tracking (VOT) in maritime environments.It addresses challenges unique to maritime scenes: such as water reflections, low-contrast objects, dynamic backgrounds, scale variation, and severe illumination changes—which are not adequately covered by generic tracking datasets.
The dataset contains 182… See the full description on the dataset page: https://huggingface.co/datasets/sivzzz/Maritime_Visual_Tracking_Dataset_MVTD.robomme-tracking-masks
RoboMME Tracking Masks
Recommended training release: releases/native_scene_v1/train/.
新版训练标签：16 个任务、1600 段、768,897 帧。原待修的 40,670 帧已全部通过既定对齐协议并合入完整 episode；另对 VideoPlaceButton ep067 的 966 帧进行了整段实例身份修正。全量独立验收 1600/1600 通过。
This release contains per-object front-camera segmentation and tracking annotations paired with the original RoboMME training images. It preserves original frame order, actions, and annotation slots. Masks come from native simulator segmentation; they are not… See the full description on the dataset page: https://huggingface.co/datasets/Tingxiang/robomme-tracking-masks.GCA_suction_franka_a01_tracking_id1_v2This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "franka",
"total_episodes": 2,
"total_frames": 215,
"total_tasks": 1,
"total_videos": 0,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 50,
"splits": {
"train": "0:2"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/hyzhang01/GCA_suction_franka_a01_tracking_id1_v2.ball-tracking
Soccer Ball Tracking Dataset
A curated dataset for detecting and tracking soccer balls in broadcast footage, specifically designed for tiny object detection challenges. This dataset supports the development of models robust to motion blur, long-shot scales, and occlusions.
Dataset Description
This public sample consists of 1,450 frames extracted from 10 video clips of professional soccer broadcasts. The data is split into two categories based on ball… See the full description on the dataset page: https://huggingface.co/datasets/infactory-ai/ball-tracking.reconstruction-tracking-syntheticGaussFly-huayiguan-tracking-100
GaussFly Huayiguan Tracking Dataset
华裔馆（huayiguan）场景的双智能体 Tracking 测试数据集。G1 在场景中导航，Rear 无人机跟踪 G1；每秒规划一次，保存 1 条最优轨迹和 10 条次优候选轨迹。
本数据集用于先验证数据加载、视觉 Tracking、候选轨迹预测/排序和训练流程能否跑通。它只包含成功 episode，不包含失败轨迹。
数据概况
项目
数值
成功 episode
100
Shard
5 个，每个 20 episode
总帧数
127,384
仿真步长
0.02 s（50 Hz）
RGB
Rear 第一视角，256 × 256 × 3，uint8
重规划记录
2,600
每次规划
11 条候选轨迹
每条候选
未来 4 s，0.1 s 间隔，共 41 个状态
自动验证
100/100 episode 通过
文件结构
README.md… See the full description on the dataset page: https://huggingface.co/datasets/jokeryao/GaussFly-huayiguan-tracking-100.Drone_Tracking_DataTracking-Any-Granularity
SAM 2++: Tracking Anything at Any Granularity
🔥 Evaluation Server | 🏠 Homepage | 📄 Paper | 🔗 GitHub
Download
We recommend using huggingface-cli to download:
pip install -U "huggingface_hub[cli]"
huggingface-cli download MCG-NJU/Tracking-Any-Granularity --repo-type dataset --local-dir ./Tracking-Any-Granularity --local-dir-use-symlinks False --max-workers 16
🔥 Latest News
[2025-11-25] Challenge Leaderboard for Tracking-Any-Granularity dataset is now… See the full description on the dataset page: https://huggingface.co/datasets/MCG-NJU/Tracking-Any-Granularity.dagger
OpenTrackVLA DAgger Recovery Dataset
Training-only DAgger recovery data generated in Habitat for OpenTrackVLA target
tracking. This release contains V6-policy failure recovery data for AT, DT, and
STT, using seeds 101 and 102.
The uploaded payload contains only the final training-ready datasets:
JSONL trajectory labels;
referenced RGB frames;
precomputed DINOv3 + SigLIP fine/coarse visual-token caches;
conversion, filtering, and integrity metadata.
Source rollouts, raw… See the full description on the dataset page: https://huggingface.co/datasets/TrackingTeam/dagger.