Datasets

Large-scale, in-the-wild robot manipulation demonstrations collected across many environments and tasks.

A growing collection and tooling ecosystem for sharing real and simulated robot learning data.

A broad collection of robot manipulation trajectories spanning tasks, objects, and environments.

A large cross-embodiment collection assembled to study generalist robot policies across institutions and hardware.

Long-horizon language-conditioned robot manipulation sequences in a simulated kitchen environment.

Robot interaction videos collected across different platforms, viewpoints, and laboratory environments.

Real robot episodes used to study large-scale, language-conditioned robotic control.

Language and video-conditioned robot demonstrations for zero-shot task generalization.

A multimodal collection of robot demonstrations designed for broad manipulation research.

Simulation environments and demonstrations for generalizable robotic manipulation skills.

Robot demonstration datasets and benchmarks for studying imitation learning methods.

A large benchmark of simulated robot manipulation tasks with demonstrations and rich observations.

Long-horizon furniture assembly tasks for evaluating real-world robot manipulation.

A benchmark of everyday household activities in interactive simulated environments.

Large-scale kitchen simulation environments and demonstrations for everyday manipulation.

A multimodal robot manipulation dataset with force, audio, vision, and action observations.

Hand-object manipulation captures with synchronized RGB-D views and pose annotations.

Egocentric RGB-D videos of human-object interactions with detailed 3D annotations.

A benchmark and demonstration suite for lifelong learning in robot manipulation.

Crowdsourced robot manipulation demonstrations collected through remote teleoperation.

A multimodal autonomous-driving dataset with synchronized cameras, lidar, radar, maps, and annotations.

Large-scale autonomous-driving data spanning perception, motion, maps, and end-to-end driving research.

A foundational driving benchmark for visual odometry, stereo, optical flow, and 3D perception.

Synthetic, photo-realistic trajectories with challenging motion and environment conditions for visual SLAM research.

Sensor, map, and motion forecasting datasets centered on autonomous driving in multiple cities.

Urban driving scenes and maps prepared for motion prediction research.

Autonomous-driving sequences with lidar, camera, semantic segmentation, and cuboid annotations.

Camera and lidar driving data with semantic labels captured across varied road conditions.

Street-scene images with fine semantic and instance annotations from cities across Europe.

A diverse driving video dataset supporting detection, lane, tracking, and segmentation tasks.

Point-wise semantic labels for the lidar sequences of the KITTI Odometry benchmark.

Repeated autonomous-vehicle traversals through Oxford across weather, traffic, and seasonal change.

Long-term driving sequences with lidar, radar, cameras, and precise localization across seasons.

Long-term campus robot traversals with lidar, cameras, odometry, GPS, and inertial sensing.

Radar and lidar sequences for place recognition and localization in changing urban environments.

Large-scale street-scene data for semantic perception, localization, and trajectory prediction.

A driving planning benchmark with real scenarios, maps, agents, and closed-loop evaluation tooling.

Large-scale autonomous-driving scenes with lidar, cameras, and 3D object annotations.

Diverse European road data for perception, localization, and autonomous-driving research.

Vehicle and roadside cooperative perception data for connected autonomous systems.

Indoor RGB-D captures with camera trajectories, reconstructed geometry, and scene annotations.

Object-centric videos with 3D bounding boxes for understanding everyday objects in real scenes.

Synchronized RGB-D sequences and ground-truth trajectories for evaluating visual odometry and SLAM.

Synchronized first-person and external-camera recordings of skilled human activities with rich annotations.

RGB-D scans of indoor environments with camera poses, surfaces, and semantic annotations.

High-fidelity indoor scene captures with DSLR images, depth, geometry, and semantic labels.

Panoramic RGB-D scans and reconstructed geometry from large indoor building environments.

High-quality reconstructions of indoor spaces for simulation and embodied AI research.

Photorealistic synthetic indoor scenes with dense geometry, material, lighting, and semantic ground truth.

A collection of high-resolution 3D spaces prepared for embodied AI navigation research.

Single-view RGB-D images of indoor scenes with 2D and 3D semantic annotations.

Indoor RGB-D video and labeled frames captured with a commodity depth camera.

Internet photo collections paired with reconstructed depth and camera geometry.

Object-centric videos with camera poses and point clouds across many everyday categories.

A large corpus of annotated 3D objects assembled for vision and generative-model research.

A deeply annotated collection of 3D CAD models organized across object categories.

Furnished indoor scenes with layouts, object models, and professionally designed room arrangements.

Egocentric sensor recordings with cameras, audio, motion, eye gaze, and spatial trajectories.

A large collection of first-person video capturing daily activities across many locations and participants.

A unified collection of human motion-capture datasets represented with a common body model.