datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
HiFi-UMI-2K
HiFi-UMI-2K: High-Fidelity Robot-Free Manipulation Data
2,000 hours released · 6 synchronized camera views · 480+ scenes · 3 mm pose accuracy · <40 µs synchronization
🌐 Project Website |
📦 Dataset |
📄 Paper: arXiv:2607.25895
Examples from the HiFi-UMI corpus. Click the image to play the video.
📚 Introduction
HiFi-UMI is a portable, high-fidelity bimanual capture system for collecting robot-free manipulation demonstrations.… See the full description on the dataset page: https://huggingface.co/datasets/simple-world-lab/HiFi-UMI-2K.World-SimReady-Home
WorldSimReady-Home
Dataset description
CAD-based SimReady assets
Optimized CAD assets with configured collision and physical properties.
Manually reviewed scenes
Physics configuration reviewed for every household scene.
Scalable task generation
Batch simulation data across robot embodiments and tasks.
WorldSimReady-Home is built from CAD-based object assets, optimized and enriched with physical properties to create simulation-ready assets. The… See the full description on the dataset page: https://huggingface.co/datasets/Yootta/World-SimReady-Home.PhysicalAI-Robotics-GR00T-Teleop-Sim
Simulation GR1 Tabletop Task 1K Dataset
Dataset Description:
The PhysicalAI-Robotics-GR00T-Teleop-GR1 dataset consists of 1000 teleoperation trajectories in simulation using the GR1 robot with upper body control. The simulation setup mimics tabletop manipulation tasks and uses RGB observations with a virtual camera. The robot is equipped with simulated Fourier hands.
This dataset is ready for non-commercial use.
Dataset Owner(s):
NVIDIA GEAR… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-GR00T-Teleop-Sim.aloha_sim_transfer_cube_humanThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "aloha",
"total_episodes": 50,
"total_frames": 20000,
"total_tasks": 1,
"total_videos": 50,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 50,
"splits": {
"train": "0:50"
},
"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/lerobot/aloha_sim_transfer_cube_human.aloha_sim_insertion_humanThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "aloha",
"total_episodes": 50,
"total_frames": 25000,
"total_tasks": 1,
"total_videos": 50,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 50,
"splits": {
"train": "0:50"
},
"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/lerobot/aloha_sim_insertion_human.aloha_sim_insertion_scriptedThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "aloha",
"total_episodes": 50,
"total_frames": 20000,
"total_tasks": 1,
"total_videos": 50,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 50,
"splits": {
"train": "0:50"
},
"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/lerobot/aloha_sim_insertion_scripted.aloha_sim_transfer_cube_scriptedThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "aloha",
"total_episodes": 50,
"total_frames": 20000,
"total_tasks": 1,
"chunks_size": 1000,
"fps": 50,
"splits": {
"train": "0:50"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4",
"features": {… See the full description on the dataset page: https://huggingface.co/datasets/lerobot/aloha_sim_transfer_cube_scripted.cobotmagic_Sim_handle_basketRAC_simcobotmagic_Sim_item_assemblycobotmagic_Sim_click_bellcobotmagic_Sim_drawer_open_placecobotmagic_Sim_mixer_operatingcobotmagic_Sim_table_rearrangementcobotmagic_Sim_water_pouringcobotmagic_Sim_manipulate_pipettedata
DualSim
Data package for the AgiBotWorld Challenge 2026 dual-arm simulation track.
Layout
dualsim/
├── eval/ # dualsim-specific evaluation set (~321 MB)
│ └── AgiBotWorldChallenge-2026-eval-dualsim/
│ ├── _lerobot/ # LeRobot v2.1 dataset (meta / data / videos / sim)
│ ├── info_dataset/ # per-episode inputs (frame.png, proprio_stats.h5, camera params, sim/)
│ ├── gt_dataset/ # per-episode ground… See the full description on the dataset page: https://huggingface.co/datasets/dual-sim/data.Mirod-Sim-3Tasks-HighQuality-3xReal-15FPS-20260914
Mirod simulation subset: three tasks, approximately 3x real frames
This release contains simulation data only, selected for mixed training with
the local Real_3tasks release. Real recordings are not included.
Folder
Task
Real reference frames
Simulation episodes
Simulation frames
Ratio
task1/dataset
Stack the small box on the other box (叠盒子)
13,785
171
41,358
3.0002
task2/dataset
Put the cup into the tray (杯子入盘)
12,361
192
37,081
2.9998
task3/dataset
Take the box… See the full description on the dataset page: https://huggingface.co/datasets/liujiting/Mirod-Sim-3Tasks-HighQuality-3xReal-15FPS-20260914.skywander-nyc-sim
SkyWander NYC Sim — Synthetic UAV Flight Corpus
2009 synthetic UAV flight clips rendered in an Unreal Engine New York City environment,
with full low-dimensional ground truth. Training/eval corpus of the SkyWander
pose-conditioned UAV world model.
Contents (per clip, keyed by uuid)
Directory
Files
What
rgb_mp4/
2009
ego-view RGB video, 1280×704
depth_mp4_gemdepth/
2009 mp4 + 2009 json
GemDepth monocular depth (gray-normalized) + per-clip scale… See the full description on the dataset page: https://huggingface.co/datasets/greenery8848/skywander-nyc-sim.dpgs_sim_faucet_5kG1_Dex1_PickPlaceCylinder_Dataset_SimThis dataset was collected using the unitree_sim_isaaclab
simulation environment, and the scene/task name used during collection is Isaac-PickPlace-Cylinder-G129-Dex1-Joint. For more details, please refer to the project's README.
usc_cloth_simThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "unknown",
"total_episodes": 1000,
"total_frames": 100000,
"total_tasks": 1,
"chunks_size": 1000,
"fps": 10,
"splits": {
"train": "0:1000"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4"… See the full description on the dataset page: https://huggingface.co/datasets/lerobot/usc_cloth_sim.robomimic_sim_square_noisyThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": null,
"total_episodes": 5000,
"total_frames": 403454,
"total_tasks": 1,
"total_videos": 10000,
"total_chunks": 5,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:5000"
},
"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/iantc104/robomimic_sim_square_noisy.aloha_sim_transfer_cube_scripted_rawinteractive-world-sim-rotate-t-data
IWS Rotate-T 90° Demonstrations (1k)
1,000 scripted expert demonstrations (+100 validation) of the rotate-the-T-90°-clockwise
task on the bimanual ALOHA push-T MuJoCo environment, in the Interactive World
Simulator (IWS) HDF5 format. Collected with
scripts/data_collection/collect_rotate_t.py from the
interactive_world_sim repo.
The T spawns upright (θ = 0) and both arms rotate it ~90° clockwise with a
closed-loop scripted policy (≤30° sub-rotations with angle feedback, so push… See the full description on the dataset page: https://huggingface.co/datasets/jacob3333/interactive-world-sim-rotate-t-data.so100_cutlery_handling_simpleThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "so100",
"total_episodes": 50,
"total_frames": 29853,
"total_tasks": 1,
"total_videos": 50,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:50"
},
"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/masato-ka/so100_cutlery_handling_simple.boris-open-base-cabinet-sim-v24This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": null,
"total_episodes": 6310,
"total_frames": 454320,
"total_tasks": 1,
"total_videos": 12620,
"total_chunks": 7,
"chunks_size": 1000,
"fps": 10,
"splits": {
"train": "0:6310"
},
"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/crislmfroes/boris-open-base-cabinet-sim-v24.KH_dt_le30_ltable_260715_1800_simThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 10,
"features": {
"action": {
"dtype": "float32",
"shape": [
19
],
"names": [
"arm_l_joint1",
"arm_l_joint2",
"arm_l_joint3",
"arm_l_joint4",
"arm_l_joint5",
"arm_l_joint6"… See the full description on the dataset page: https://huggingface.co/datasets/shju0519-hf/KH_dt_le30_ltable_260715_1800_sim.single_pickplace
pick_and_place-300 — Unitree G1 + Dex3, "pick octopus and place inside brown basket"
LeRobot v2.1 dataset. Teleoperated bimanual G1 with Dex3 hands; lower body under a GR00T
whole-body-control policy, upper body teleoperated.
Episodes
349 (322 positive demos + 27 negative samples)
Frames
161,440 (2.24 h @ 20 fps)
FPS
20
Cameras
3 × h264 640×480 yuv420p
State / action
43-dim whole body (float64)
Task string
pick octopus and place inside brown basket
Size… See the full description on the dataset page: https://huggingface.co/datasets/simpk/single_pickplace.av_aloha_sim_slot_insertionThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": null,
"total_episodes": 100,
"total_frames": 17741,
"total_tasks": 1,
"total_videos": 600,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 25,
"splits": {
"train": "0:100"
},
"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/iantc104/av_aloha_sim_slot_insertion.