Source Videos for OneStreamer-1M Dataset
English and Chinese video files for Visual Question Answering tasks from the OneStreamer-Core project.
- Visual question answering
- English
- Chinese
Hugging Face datasets with plain-language descriptions: what's inside and what models to train on them. 4,788 in the catalog so far, with new ones added every hour.
Answered by Gemini 4 · the chat is kept only in this tab
English and Chinese video files for Visual Question Answering tasks from the OneStreamer-Core project.
1,274 hours of first-person video with synchronized 7-point IMU hand tracking during household chores.
A large-scale video dataset in MOTChallenge format for multi-object tracking in dynamic sports scenes.
6.18 million RGB+Depth frames and over a million text instructions for training dual-arm robotic manipulators.
A large corpus of video and telemetry for the KUKA iiwa manipulator to train robot control models.
Large collection of video clips, captions, and spatial reasoning tasks for training multimodal models.
Multimodal dataset featuring optical motion capture and egocentric video for humanoid robot learning.
Over 215 hours of recordings from 41 cameras and tactile grasps for studying the deformation of 198 household objects upon contact.
Massive, regularly updated scrape of public videos, comments, and stats from the Chinese platform Bilibili.
Video and tabular robot teleoperation data for the DreamDojo world model.
Video recordings and real robot action trajectories for offline-to-online RL research.
Teleoperation recordings of humanoid robots in household tasks for training embodied AI systems.
192 hours of video and telemetry of bimanual robot manipulation for training control systems.
A multimodal dataset with 3D data, images, and videos of robot motions, including stair climbing and descending.
106 hours of free-camera flights across 17 games for 3D reconstruction and video generation.
280+ hours of synchronized video from 1,800 cameras for tracking and 2D/3D detection.
12.2K episodes of real-world dual-arm robot teleoperation for training VLA-style models.
20 thousand episodes of executing 100 household tasks in LeRobotDataset v3 format for training robotics models.
About 123,000 fully annotated synthetic video clips covering warehouse safety scenarios.
Over one million frames and 34 thousand video recordings of 24 tasks performed by a Franka robot in LeRobot format.
10,000 hours of real human work video with head and hand pose annotations in video format for robotics.
Bird's-eye view frames from the AirSim simulator for depth estimation and segmentation.
1,000 simulations to verify how accurately and safely robots perform household tasks.
Humanoid robot simulation demonstrations for training VLA policies in LeRobot format.