datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
test-videosRepaired_videosstatic_videosremake_videoshyz-sample-videos
hyz-sample-videos
Part of the TEKNOFEST Aviation & UAV AI Challenge Datasets (UYZ / HYZ) collection.
Standardized competition, annotation, and evaluation benchmarks for TEKNOFEST Ulaşımda Yapay Zeka (UYZ) and Havacılıkta Yapay Zeka (HYZ) drone perception challenges organized by the Turkish Technology Team (T3) Foundation and TÜBİTAK.
Original Source & License
Official Source / Project: TEKNOFEST Aviation & UAV AI Challenge Datasets (UYZ / HYZ)
License: Creative… See the full description on the dataset page: https://huggingface.co/datasets/huseyincavus/hyz-sample-videos.ssim-reference-videoswat.ai-so101-videosgo2_rl_gym_videos
go2 rl gym videos
This is all original videos for project:
Code:
Train: go2_rl_gym
Evaluation: RoboGauge
Deploy: unitree_cpp_deploy
Page: robogauge.github.io
Paper: Toward Reliable Sim-to-Real Predictability for MoE-based Robust Quadrupedal Locomotion
VideoScore-Bench
Overview
VideoFeedback-Bench is derived from four benchmarks or datasets: VideoFeedback, EvalCrafter, GenAI-Bench and VBench.
Examples
Citation
@article{he2024videoscore,
title = {VideoScore: Building Automatic Metrics to Simulate Fine-grained Human Feedback for Video Generation},
author = {He, Xuan and Jiang, Dongfu and Zhang, Ge and Ku, Max and Soni, Achint and Siu, Sherman and Chen, Haonan and Chandra, Abhranil and Jiang, Ziyan and Arulraj, Aaran and… See the full description on the dataset page: https://huggingface.co/datasets/TIGER-Lab/VideoScore-Bench.road-traffic-in-serbia-videos-and-images
Vehicle Detection Dataset
Dataset comprises 5,000 high-resolution (≥1080p) videos of Serbian road traffic captured from bridge vantage points, accompanied by annotated frame images with JSON bounding box annotations. It is designed for research in vehicle detection and classification, focusing on analyzing traffic flows and traffic density to support AI-driven traffic management solutions.
By utilizing this dataset, researchers and developers can enhance their capabilities in… See the full description on the dataset page: https://huggingface.co/datasets/UniDataPro/road-traffic-in-serbia-videos-and-images.videosFANVID-Face_and_License_Plate_Recognition_in_Low-Resolution_Videos
FANVID: Face and License Plate Recognition in Low-Resolution Videos
Overview
FANVID is a benchmark dataset designed to advance research in face detection and matching and license plate recognition under challenging low-resolution surveillance video conditions. Unlike existing datasets, FANVID features faces and license plates that are unrecognizable in individual frames, encouraging models to leverage temporal context across video sequences for improved recognition.… See the full description on the dataset page: https://huggingface.co/datasets/kv1388/FANVID-Face_and_License_Plate_Recognition_in_Low-Resolution_Videos.MOSS-VL-Demo-Videossample_factory_videosT2V-CompBench-VideosDF40-Fake-VideosVideoSSR-30kvideo-std-manip
Video Standard Manipulation Dataset
This dataset is a collection of simple and traditional localized video manipulations, such as: splicing, color correction, contrast enhancement, bluring, and noise addition. The dataset is designed to be used for training and evaluating video manipulation detection models. We used this dataset to train the VideoFACT model, which is a deep learning model that uses attention, scene context, and forensic traces to detect a wide variety of video… See the full description on the dataset page: https://huggingface.co/datasets/ductai199x/video-std-manip.text2video-energy-benchmark-generated-videosvideosubtitlelium-videos
Lium short videos
Short product videos from Lium, the GPU rental marketplace. They're made by code from live lium.io prices and real screen recordings. Prices on screen were live when each video was made; providers set prices, so check lium.io/pricing for the current ones.
File
What it shows
Recorded
lium-video-01-b300.mp4
8x NVIDIA B300 from one CLI command, $68 an hour
prices from 24 Sep 2026 02:29 UTC
lium-video-02-agents.mp4
An AI agent finding, renting and… See the full description on the dataset page: https://huggingface.co/datasets/gpu-rentals/lium-videos.SplineCoT-LIBERO-videosBenchmark_Videosrobomme_gaze_eval_videos
RoboMME eval videos — GR00T N1.6 gaze all-past
Simulator rollout videos on the RoboMME benchmark (test split, 50 episodes/task, max 1500 steps, n_action_steps=16).
Structure: <model>/<task>/<video>.mp4 + per-task simulation_results.csv (has per-episode task_instruction) and README.md.
Models
adaptive_ns16/ — gaze all-past, adaptive, checkpoint-120000.
tl0.7_ns16/ — gaze all-past, tl0.7, checkpoint-120000.
tl0.8_ns16/ — gaze all-past, target_loss 0.8, budget 2048… See the full description on the dataset page: https://huggingface.co/datasets/hmkang/robomme_gaze_eval_videos.robotwin-da3-contrast-eval-videos
RoboTwin DA3 contrast eval videos
Clean-phase rollouts from the 50-task RoboTwin evals:
DA3 off: FasterWAM M1 step_029355 (job 3608)
With DA3: frozen-WAM + spatial bank step_015936 (job 3260)
Each task has two videos from the same clean seed: DA3 succeeds, DA3-off fails.
4x-speed MP4s are in videos_4x/ (all eight clips in one folder; task name is in the filename).
task
seed
DA3-off (fail)
with DA3 (success)
move_can_pot
4300017
episode 15
episode 16
stamp_seal… See the full description on the dataset page: https://huggingface.co/datasets/mah-asd/robotwin-da3-contrast-eval-videos.Anti-Spoofing-Real-Videos
Face Anti Spoofing Dataset - 98 000+ files
Dataset features 98,000+ files of real photos and videos of people from 170+ countries, representing 70,000+ unique individuals. By leveraging this dataset, developers can enhance spoofing detection techniques, improve recognition systems, and deploy anti-spoofing algorithms capable of preventing fraud in deep learning-based solutions.- Get the data
Dataset characteristics:
Characteristic
Data
Description
Live… See the full description on the dataset page: https://huggingface.co/datasets/ud-biometrics/Anti-Spoofing-Real-Videos.1k-ranked-videos-coherence
1k Ranked Videos
This dataset contains approximately one thousand videos, ranked from most preferred to least preferred based on human feedback from over 25k pairwise comparisons. The videos are rated solely on coherence as evaluated by human annotators, without considering the specific prompt used for generation. Each video is associated with the model name that generated it.
The videos are sampled from our benchmark dataset text-2-video-human-preferences-pika2.2. Follow us to… See the full description on the dataset page: https://huggingface.co/datasets/Rapidata/1k-ranked-videos-coherence.HUI360-Videos
HUI360-Videos
HUI360: A 360° Egocentric Dataset and Baselines for Human-Robot Interaction Anticipation (IEEE FG 2026)
Video data for HUI360, a large-scale 360° egocentric dataset for human-robot interaction anticipation in the wild.
Related resources
Resource
Link
Project website
hucebot.github.io/hui360
Paper (HAL)
hal.science/view/index/docid/5609928
Annotation & processing pipeline
Interact360
Skeleton annotations (open)… See the full description on the dataset page: https://huggingface.co/datasets/rlorlou/HUI360-Videos.spinning-top-simulation-1.5k-videos-v2sifanshang-cat-videos
橘猫博士 · AI 推广短视频数据集
用 AI 自动生成的「橘猫博士」防尘/防静电服装推广短视频成片。
这是 GitHub 工程 ai-contents-av-gen 的视频素材部分——
代码和配置在 GitHub，视频成片托管在这里。
📥 怎么下载
只要视频： 点上方 Files → 进 03_工作结果/待发布/，每个 .mp4 旁有下载按钮。
命令行批量拉取：
pip install -U "huggingface_hub[cli]" # 或 brew install huggingface-cli
# 拉全部成片
hf download liu1/sifanshang-cat-videos --repo-type dataset \
--include "03_工作结果/*" --local-dir .
公开数据集，下载无需登录、无需 token。
📂 目录
目录
内容
03_工作结果/待发布/
国内版成片
03_工作结果/待发布_海外/… See the full description on the dataset page: https://huggingface.co/datasets/liu1/sifanshang-cat-videos.