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.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.her-trees-simple-we
Her Trees Simple World-Engine Dataset
Synthetic 832x480, 81-frame solution videos for the Her Trees letter-assembly task.
Train: 3,000 cases; easy, medium, and hard have 1,000 each.
IID validation: validation/; 25 cases per difficulty and 75 total.
Same-letter fragments share one color and move independently.
See SPLITS.md and the split manifests for seeds and validation details.
directional_pick_place_simpledesk_diverse_targets_tested40_200demos_20260716_molmobot
Directional SimpleDesk Pick-and-Place Diverse Targets
This dataset contains 200 successful scripted Franka demonstrations converted
from native MolmoSpaces output into the MolmoBot/Synthmanip training layout.
The task is to pick up one tabletop object and place it either to the left of
or to the right of a second object, from the robot's point of view.
The dataset is balanced by direction: 100 demonstrations use left prompts and
100 use right prompts. It covers 40 fixed initial… See the full description on the dataset page: https://huggingface.co/datasets/ccwatson/directional_pick_place_simpledesk_diverse_targets_tested40_200demos_20260716_molmobot.hepha_act_100_simple_drawer_5
tmeynier/hepha_act_100_simple_drawer_5
LeRobot-style behavior-cloning dataset generated from the Hepha MuJoCo simulation.
Summary
Robot type: hepha_mujoco
Codebase version: v3.0
Episodes: 100
Frames: 200000
FPS: 30
Joint normalization: min_max_0_1
Features
timestamp: float32 [1]
frame_index: int64 [1]
episode_index: int64 [1]
index: int64 [1]
task_index: int64 [1]
episode.drawer_index: int64 [1]
episode.cube_position: float32 [3]… See the full description on the dataset page: https://huggingface.co/datasets/tmeynier/hepha_act_100_simple_drawer_5.octo-small-simpler-cube-stack-rollout-bank-50
Octo-Small SIMPLER cube-stack rollout bank
This bank contains exactly 50 deterministic Octo-Small
rollouts for StackGreenCubeOnYellowCubeBakedTexInScene-v1: 3
successes and 47 failures. Every episode has a complete
61-frame H.264 video, a seven-frame contact sheet, losslessly stored per-step
telemetry, derived phase/geometry metrics, an external-LLM diagnosis, exact
evidence values, and a future action-patch hypothesis for failures.
Dataset:… See the full description on the dataset page: https://huggingface.co/datasets/lsnu/octo-small-simpler-cube-stack-rollout-bank-50.sort_b601_simpleThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_yaw.pos",
"wrist_roll.pos",
"gripper.pos"
]… See the full description on the dataset page: https://huggingface.co/datasets/adrfm/sort_b601_simple.sort_b601_simple_makelabThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_yaw.pos",
"wrist_roll.pos",
"gripper.pos"
]… See the full description on the dataset page: https://huggingface.co/datasets/adrfm/sort_b601_simple_makelab.rollout_sort_b601_simple_filtered_act_v2_changed_envThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_yaw.pos",
"wrist_roll.pos",
"gripper.pos"
]… See the full description on the dataset page: https://huggingface.co/datasets/adrfm/rollout_sort_b601_simple_filtered_act_v2_changed_env.koch_pnp_simple_50This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "koch",
"total_episodes": 50,
"total_frames": 13101,
"total_tasks": 1,
"total_videos": 100,
"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/hannesill/koch_pnp_simple_50.rollout_sort_b601_simple_filtered_act_v2This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_yaw.pos",
"wrist_roll.pos",
"gripper.pos"
]… See the full description on the dataset page: https://huggingface.co/datasets/adrfm/rollout_sort_b601_simple_filtered_act_v2.hepha_act_200_simple_drawer_5
tmeynier/hepha_act_200_simple_drawer_5
LeRobot-style behavior-cloning dataset generated from the Hepha MuJoCo simulation.
Summary
Robot type: hepha_mujoco
Codebase version: v3.0
Episodes: 200
Frames: 253867
FPS: 30
Joint normalization: min_max_0_1
Features
timestamp: float32 [1]
frame_index: int64 [1]
episode_index: int64 [1]
index: int64 [1]
task_index: int64 [1]
episode.drawer_index: int64 [1]
episode.cube_position: float32 [3]… See the full description on the dataset page: https://huggingface.co/datasets/tmeynier/hepha_act_200_simple_drawer_5.pensInHolder-simple-localThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "so101_follower",
"total_episodes": 1,
"total_frames": 899,
"total_tasks": 1,
"total_videos": 2,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:1"
},
"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/sebastiandavidlee/pensInHolder-simple-local.simple_2d_dataset_lv2This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": null,
"total_episodes": 1000,
"total_frames": 32099,
"total_tasks": 1,
"total_videos": 1000,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 10,
"splits": {
"train": "0:1000"
},
"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/seann999/simple_2d_dataset_lv2.sort_b601_simple_filteredThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_yaw.pos",
"wrist_roll.pos",
"gripper.pos"
]… See the full description on the dataset page: https://huggingface.co/datasets/adrfm/sort_b601_simple_filtered.sort_b601_simple_makelab_diepenbeekThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_yaw.pos",
"wrist_roll.pos",
"gripper.pos"
]… See the full description on the dataset page: https://huggingface.co/datasets/adrfm/sort_b601_simple_makelab_diepenbeek.fixture-experiment-prism-simple-50_20260902_214241This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos",
"gripper.pos"
],
"shape": [
6… See the full description on the dataset page: https://huggingface.co/datasets/filesmuggler/fixture-experiment-prism-simple-50_20260902_214241.simple-lego-pickup-mono-2
simple-lego-pickup-mono-2
This dataset was generated using a phospho dev kit.
This dataset contains a series of episodes recorded with a robot and multiple cameras. It can be directly used to train a policy using imitation learning. It's compatible with LeRobot and RLDS.
sort_b601_simple_20260828_104340This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_yaw.pos",
"wrist_roll.pos",
"gripper.pos"
]… See the full description on the dataset page: https://huggingface.co/datasets/adrfm/sort_b601_simple_20260828_104340.pensInHolder-simple-one-two
pensInHolder-simple-one-two
This dataset was generated using a phospho starter pack.
This dataset contains a series of episodes recorded with a robot and multiple cameras. It can be directly used to train a policy using imitation learning. It's compatible with LeRobot and RLDS.
so100_jenga_box_simpleThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "so100",
"total_episodes": 5,
"total_frames": 1711,
"total_tasks": 1,
"total_videos": 5,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:5"
},
"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/asafxrev/so100_jenga_box_simple.simple-lego-pickup-mono-train
simple-lego-pickup-mono-2
This dataset was generated using a phospho dev kit.
This dataset contains a series of episodes recorded with a robot and multiple cameras. It can be directly used to train a policy using imitation learning. It's compatible with LeRobot and RLDS.
mecharm_simple_pnpThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"joint_1.pos",
"joint_2.pos",
"joint_3.pos",
"joint_4.pos",
"joint_5.pos",
"joint_6.pos",
"gripper.pos"
],
"shape": [… See the full description on the dataset page: https://huggingface.co/datasets/Lanuchi/mecharm_simple_pnp.franka_pick_and_place_avoid_laptop_simple10_randomized_all_easierline_300x_seed20260624_molmobotsimple_2d_datasetThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": null,
"total_episodes": 50,
"total_frames": 1303,
"total_tasks": 1,
"total_videos": 50,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 10,
"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/seann999/simple_2d_dataset.mimic_simpleClothFolding_v1This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "mimic_follower",
"total_episodes": 222,
"total_frames": 384490,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:222"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/Mimic-Robotics/mimic_simpleClothFolding_v1.PND_Adam-U_pick-simpleThis data is part of the training data for Being-H0.5, produced by BeingBeyond.
License
This dataset is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0).
Citation
Being-H0.5
@article{beingbeyond2026beingh05,
title={Being-H0.5: Scaling Human-Centric Robot Learning for Cross-Embodiment Generalization},
author={Luo, Hao and Wang, Ye and Zhang, Wanpeng and Zheng, Sipeng and Xi, Ziheng and Xu, Chaoyi and Xu, Haiweng and Yuan… See the full description on the dataset page: https://huggingface.co/datasets/BeingBeyond/PND_Adam-U_pick-simple.minecraft_simplesimple_dataset_pick_and_place_distractorsThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "so100_follower",
"total_episodes": 37,
"total_frames": 20732,
"total_tasks": 1,
"total_videos": 74,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:37"
},
"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/Clementppr/simple_dataset_pick_and_place_distractors.babybel_simple2