232 dated dataset releases

Dataset releases.

Browse published dataset records with source-backed, day-level release dates. Results stay pinned to one Catalog generation as you move between pages.

232 records

Source-backed datasets.

Sort

Loading source media

Dataset Series

Open

EgoSuite-Open100K

Lightwheel · 2026-08-07

EgoSuite-Open100K is a large egocentric corpus of people performing everyday work tasks, captured on a head rig with stereo cameras and — on the EgoPro line — a camera on each wrist. Recordings carry 3D hand pose, full-body pose on the body subsets, and subtask-level semantic segments, and ship in LeRobot v3 and MCAP. Read the hours claim carefully. The…

Egocentriccustomlerobot

Catalog score

362K episodes

Readiness 73 · verified 2026-08-22

ArmnetBench v0.1 Dataset

Armnet · 2026-07-27

ArmnetBench v0.1 is a real-world manipulation benchmark: 7 policies run on a fleet of low-cost SO-101 cells, with every episode human-scored. It is released as evaluation data rather than a training corpus, which makes it unusual here — most of the catalog is demonstrations, and this is what happens when policies actually try. What is in it. 3,718 episodes…

ManipulationApache-2.0lerobot

Catalog score

3,718 episodes

Readiness 70 · verified 2026-09-08

RoboDojo

MMLab@HKU and collaborators · 2026-07-05

RoboDojo is a sim-and-real evaluation benchmark built to answer a blunt question: how good are generalist manipulation policies really? The answer it reports is sobering — the best of 30 policies evaluated in simulation reached only 8.80% success against 76.03% for human teleoperation. 60 tasks total: 42 simulation tasks (NVIDIA Isaac Sim 5.1 / Isaac Lab…

ManipulationApache-2.0hdf5

Catalog score

5,300 episodes

Readiness 77 · verified 2026-09-08

HABIT (Human Aware Behavior and Interaction Training dataset)

Config · 2026-06-30

HABIT (Human-Aware Behavior and Interaction Training dataset) is a large-scale robot manipulation dataset collected in human-present environments, where a co-present human shares the workspace and interacts with the robot in every episode. 10,563 episodes, 164 hours of bimanual manipulation across 60 tasks Organized into three human-robot interaction roles…

ManipulationCC-BY-4.0custom

Catalog score

10.6K episodes

Readiness 73 · verified 2026-09-08

ABC-130k

UC Berkeley / MIT / Amazon FAR / CMU / XDOF · 2026-06-25

ABC-130k is an open bimanual robot manipulation dataset released with the paper 'Scalable Behavior Cloning with Open Data, Training, and Evaluation'. It contains roughly 130,919 episodes (43,090 annotated) totaling about 3,598 hours collected on bimanual YAM stations with two 6-DoF arms and parallel-jaw grippers, spanning ~195 tasks such as pick-and-place,…

ManipulationApache-2.0custom

Catalog score

130.9K episodes

Readiness 73 · verified 2026-09-08

HIW-500: Humanoids In-the-Wild

BitRobot · 2026-06-25

HIW-500 (Humanoids In-the-Wild) is a large-scale dataset for whole-body humanoid robot learning in natural home environments, capturing human teleoperation demonstrations on the Unitree G1 across 12 real homes in Southeast Asia. It contains 500+ hours and 23K+ episodes spanning 10+ household tasks with 161 subtask labels (~10 TB). Each episode records…

Manipulationunknownrosbag

Catalog score

23K episodes

Readiness 77 · verified 2026-09-09

LeRobot Shirt-Folding Dataset

LeRobot (Hugging Face) · 2026-06-22

The shirt-folding dataset from LeRobot's open-source 'Unfolding Robotics: Open-Source Shirt Folding from Data to Deployment' project, which trained a bimanual robot to fold t-shirts at a 90% success rate. Collected via real-world leader-follower teleoperation across 8 bimanual setups optimizing for diversity (25+ t-shirts, 8 backgrounds, varied camera/robot…

ManipulationApache-2.0lerobot

Catalog score

5,688 episodes

Readiness 70 · verified 2026-09-09

EgoInfinity

Rice RobotPI Lab · 2026-06-16

EgoInfinity is not a static dataset but a modular, upgradable data engine: it takes ordinary internet video of people manipulating objects and automatically recovers metric 4D hand–object interaction, then retargets that motion into executable robot joint trajectories — no manual annotation and no depth sensors. Runs over Meta FAIR's Action100M corpus —…

EgocentriccustomVideo / RGB

Catalog score

1,150 episodes

Readiness 77 · verified 2026-09-08

LESS (Local Encoder for Spatial Sensing)

Technion – Israel Institute of Technology · 2026-06-12

LESS ("More with LESS: Local Scene Representations for Tactile Imaging", RSS 2026) tackles artificial palpation — reconstructing what is inside a soft object purely by touching it. The motivating application is breast cancer screening, where over 40% of cases are first detected by palpation. >800 hours of robotic tactile interaction, released as ~66 GB…

ManipulationCC-BY-4.0custom

Catalog score

800 hours

Readiness 70 · verified 2026-09-09

FTP-1 Dataset

FTP-1 (Yuan et al.) · 2026-06-11

FTP-1-Dataset is the heterogeneous pretraining corpus behind FTP-1 (A Generalist Foundation Tactile Policy Across Tactile Sensors for Contact-Rich Manipulation). It aggregates roughly 3,000 hours of contact-rich manipulation data from 26 sources spanning human demonstrations, dexterous-hand, and gripper robots, recorded with 21 distinct tactile sensors…

ManipulationMITzarr

Catalog score

3,000 hours

Readiness 73 · verified 2026-09-08

Grid Egocentric Residential Samples

Hugging Face · aryan56789gridai/grid-egocentric-residential-samples · 2026-05-28

A CC-BY-4.0 household egocentric sample set tagged for manipulation, VLA, imitation learning, and robot learning.

EgocentricCC-BY-4.0video

Catalog score

<1K rows

Readiness 69 · verified 2026-07-28

Industrial Workplace Egocentric FHD Samples

Hugging Face · TrainThemAI/Industrial-Workplace-Egocentric-FHD-Samples · 2026-05-24

A small MIT-licensed set of industrial and workplace egocentric video samples for robotics, VLA, imitation learning, and manipulation demos.

EgocentricMITvideo

Catalog score

<1K rows

Readiness 69 · verified 2026-07-28

Wearable AI

Hugging Face · facebook/wearable-ai · 2026-05-15

A benchmark for long-form QA, conversational QA, and proactive assistant behavior over first-person wearable-camera videos.

EgocentricMITJSONL

Catalog score

2,100 clips

Readiness 69 · verified 2026-07-28

EgoStation GoPro Pick-and-Place

Hugging Face · zeno-labs/egostation-gopro-pick-and-place-v1 · 2026-05-12

A non-gated CC-BY-NC-4.0 LeRobot-style dataset with GoPro first-person manipulation, hand-world coordinates, and 6DoF trajectory tags.

EgocentricCC-BY-NCLeRobot

Catalog score

10K-100K rows

Readiness 57 · verified 2026-07-28

EgoStation Catalog

Hugging Face · zeno-labs/egostation-catalog-v1 · 2026-05-11

A CC-BY-NC-4.0 catalog dataset for egocentric first-person manipulation data, including GoPro pick-and-place metadata.

EgocentricCC-BY-NCParquet

Catalog score

<1K rows

Readiness 56 · verified 2026-07-28

EgoMM

Hugging Face · ldkong/EgoMM · 2026-05-10

A tri-modal egocentric dataset built from EgoLife and Ego-Exo4D sources, packaged into fixed-duration clips with optional narrations.

EgocentricApache-2.0MP4

Catalog score

55.5K clips

Readiness 70 · verified 2026-07-28

MolmoAct Dataset

Allen Institute for AI (Ai2) · 2026-05-04

Ai2's self-collected MolmoAct real-robot manipulation dataset, used to train MolmoAct action reasoning models, covering household and tabletop domains. The Household split has 7,529 episodes across 115 tasks with varying viewpoints; the Tabletop split adds 2,959 episodes across 19 tasks with a fixed exocentric camera (~10,488 episodes total). It uses a…

ManipulationApache-2.0lerobot

Catalog score

10.5K episodes

Readiness 73 · verified 2026-09-08

MolmoAct2 Bimanual YAM Dataset

Allen Institute for AI (Ai2) · 2026-05-04

A large-scale bimanual manipulation dataset collected by Ai2 for MolmoAct2 using paired YAM follower arms. It contains 32,246 episodes (76M frames) spanning 34 tabletop tasks with 720+ hours of teleoperated demonstrations. Each episode provides three RGB camera views (left, right, top) plus 14-dim joint/gripper states, in LeRobot format with per-episode…

ManipulationApache-2.0lerobot

Catalog score

32.2K episodes

Readiness 73 · verified 2026-09-09

MolmoAct2 SO-100/SO-101 Dataset

Allen Institute for AI (Ai2) · 2026-05-04

A MolmoAct2 resource from Ai2 providing per-episode annotated language instructions for low-cost SO-100 and SO-101 arm data sourced from 1,220 community LeRobot repositories (377 users). It does not re-host trajectories; each source dataset gets a tasksannotated.parquet of refined instructions indexed by episode, with the source repos listed in…

ManipulationApache-2.0lerobot

Catalog score

3.46 MB size

Readiness 70 · verified 2026-09-09

Egocentric Adjust Bottle

Hugging Face · SUZ-tsinghua/egocentric_adjust_bottle · 2026-05-02

A LeRobot-style robotics dataset for an egocentric bottle adjustment task, packaged as Parquet with image, text, and timeseries modalities.

EgocentricApache-2.0LeRobot

Catalog score

10K-100K rows

Readiness 73 · verified 2026-07-28

Egocentric Activity Sample

Hugging Face · WhissleAI/egocentric-activity-sample · 2026-04-25

A CC-BY-4.0 sample dataset for egocentric activity recognition, visual QA, temporal localization, narrations, hands, and objects.

EgocentricCC-BY-4.0Parquet

Catalog score

<1K rows

Readiness 70 · verified 2026-07-28

LIBERO-Long LeRobot

Robbyant / LIBERO · 2026-04-24

A LeRobot-formatted LIBERO-Long release used for post-training and evaluating video-action and VLA policies on extended tasks.

ManipulationLeRobotActions

Catalog score

Post-training role

Readiness 73 · verified 2026-07-28

EgoWorld

Hugging Face · haoyang-li/EgoWorld · 2026-03-30

A compact egocentric bimanual manipulation dataset with world-frame 3D hand poses, MANO meshes, camera trajectories, depth maps, and 40D action/state vectors.

EgocentricCC-BY-NCLeRobot v3

Catalog score

2 episodes

Readiness 63 · verified 2026-07-28

ManipArena

ManipArena (SYSU x MBZUAI) · 2026-03-30

ManipArena is a comprehensive real-world benchmark for evaluating reasoning-oriented generalist robot manipulation, run as a CVPR 2026 Embodied AI Workshop challenge. It provides 10,812 expert trajectories (~188 hours, 13.5M frames) across 20 real tasks spanning execution-reasoning, semantic-reasoning, and long-horizon mobile manipulation, plus 3…

Mobile robotsApache-2.0lerobot

Catalog score

10.8K episodes

Readiness 73 · verified 2026-09-08

View record