datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
PhysicalAI-Robotics-Manipulation-Kitchen-Demos
PhysicalAI-Robotics-Manipulation-Kitchen-Demos
We provide a 600 hours of human-teleoperated demonstrations across 316 different tasks, totalling 55k trajectories.
The datasets are collected using Franka Panda robot with an Omron mobile base.
The datasets follow the LeRobot format. Here is an overview of important elements of each dataset:
Click to expand dataset structure
lerobot/
├── meta/ # Metadata files describing the dataset
│ ├── info.json… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-Manipulation-Kitchen-Demos.PhysicalAI-Robotics-Manipulation-SingleArm
Dataset Description:
PhysicalAI-Robotics-Manipulation-SingeArm is a collection of datasets of automatic generated motions of a Franka Panda robot performing operations such as block stacking, opening cabinets and drawers. The dataset was generated in IsaacSim leveraging task and motion planning algorithms to find solutions to the tasks automatically [1, 3]. The environments are table-top scenes where the object layouts and asset textures are procedurally generated [2].This dataset… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-Manipulation-SingleArm.PhysicalAI-Robotics-Manipulation-Objects-Kitchen-MJCF
Manipulation Objects Kitchen MJCF
We provide a collection of digital 3D mjcf assets intended for use in a simulated kitchen environment.
The assets are broadly divided into 2 categories: fixtures and objects. The fixture assets are comprised of interactable kitchen appliances such as stoves, microwaves, and ovens. The object assets consist of common kitchen objects such as saucepans and glass cups.
Preview
Objects
Fixtures
Objects… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-Manipulation-Objects-Kitchen-MJCF.functional-manipulation-benchmark
Functional Manipulation Benchmark
This robot learning dataset is a part of the paper "FMB: a Functional Manipulation Benchmark for Generalizable Robotic Learning". It includes 22,550 expert
demonstration trajectories across different skills required to solve the Single-Object and Multi-Object Manipulation Tasks presented in the paper.
Link to paper: https://arxiv.org/abs/2401.08553
Link to website: https://functional-manipulation-benchmark.github.io
Dataset Structure… See the full description on the dataset page: https://huggingface.co/datasets/charlesxu0124/functional-manipulation-benchmark.kitchen-workspace-understanding-safe-manipulation
Kitchen Workspace Understanding & Safe Manipulation
Generated by datapack-import.ts
This dataset mirrors public data-pack render outputs from Physicl.
Each row represents one render view. The image column contains a stable URL to the primary render image uploaded under /data; image_path stores the relative repository path and data_commit_sha pins the Hugging Face dataset commit used by those URLs. Files are uploaded as downloaded unless optional PNG recompression is enabled by… See the full description on the dataset page: https://huggingface.co/datasets/physicl/kitchen-workspace-understanding-safe-manipulation.PhysicalAI-Robotics-Manipulation-Augmented
Dataset Description:
This is a fully annotated, synthetically generated dataset consisting of 1,000 demonstrations of a single Franka Panda robot arm performing a fixed-order three-cube stacking task in Isaac Lab. The robot consistently stacks cubes in the order: blue (bottom) → red (middle) → green (top).
The dataset was produced using the following pipeline:
Collected 10 human teleoperation demonstrations of the stacking task.
Used Isaac Lab’s Mimic tool [1] to simulate 1,000… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-Manipulation-Augmented.zerograv-manipulation-trajectories-state-v1
Zero-Gravity Manipulation Trajectories (State-Only, v1) Dataset Card
Fourth collection round on the DROID rig, superseding
stray-light/zerograv-manipulation-trajectories-state-v0 —
same general recipe (state-only, ~1000 episodes/task, QA-filtered), but with a
recorded-action bug fixed, hold-segment trimming applied, corrected QA filtering
per task, and one new task (PickSingleYCB). Full per-task hyperparameters below,
sourced directly from each training run's own logged… See the full description on the dataset page: https://huggingface.co/datasets/stray-light/zerograv-manipulation-trajectories-state-v1.everyday-manipulation-3d-raw
Everyday Manipulation 3D (raw RGB-D)
1,513 clips · 10.28 hours · 279 GiB · 4 participants · 10 manipulation tasks · 42 recording sittings
Chest-mounted iPhone Pro capture of everyday two-handed manipulation by
CaryX AI. Clips were recorded with
Record3D, an iOS app that captures the
iPhone's LiDAR RGB-D stream. Each clip is the app's .r3d recording with the
audio track removed; the sensor streams are unmodified: synchronised RGB,
metric LiDAR depth, per-frame ARKit 6-DoF camera… See the full description on the dataset page: https://huggingface.co/datasets/CaryxAI/everyday-manipulation-3d-raw.zerograv-manipulation-regular-gravity-trajectories-state-v1
Regular-Gravity (Vanilla-Matched) Manipulation Trajectories (State-Only, v1) Dataset Card
The regular-gravity control arm for
stray-light/zerograv-manipulation-trajectories-state-v1.
Same DROID rig, same tasks, same PPO recipe, same per-task QA filters, same
hold-segment trimming, same LeRobot layout, and ~1000 episodes per task — the
only intended difference is the physics: gravity is always on, whereas the
zero-gravity (ZG) datasets float the object (fully, or until… See the full description on the dataset page: https://huggingface.co/datasets/stray-light/zerograv-manipulation-regular-gravity-trajectories-state-v1.Arena-G1-Loco-Manipulation-Task
Dataset Description:
The Arena-G1-Loco-Manipulation-Task dataset is multimodal collections of trajectories generated in Isaac Lab. It supports humanoid (G1) loco-manipulation task in IsaacLab-Arena environment. Each entry provides the full context (state, vision, language, action) needed to train and evaluate generalist robot policies for box pick and place task.
Dataset Name
# Trajectories
G1 Loco-Manipulation Task
50
This dataset is ideal for behavior cloning… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Arena-G1-Loco-Manipulation-Task.warehouse-manipulation-and-states
Warehouse Manipulation & States
Try it before you download. Every item in this dataset is live in
projectsim Lab: drag the box flaps, roll the cage casters, then copy the API call that generates a set like it.
Jump straight to a class: Boxes · Cages · Pallet Boxes.
Procedurally generated, sim-ready 3D warehouse assets - cardboard boxes,
pallet-and-box stacks, and wire logistics cages - shipped as OpenUSD (.usda)
and MuJoCo XML (.xml, MJCF) with per-zone PBR maps… See the full description on the dataset page: https://huggingface.co/datasets/projectsim/warehouse-manipulation-and-states.ego-tactile-manipulation
Ego-Tactile Manipulation
Egocentric video + dense two-hand tactile + touch-grounded action labels - by OpenGraph Labs.
Four episodes visualized in our dashboard - egocentric video with live tactile & sensor signals.
Synchronized ego + touch human-manipulation data is rare. This is a clean 1.28-hour sample from OpenGraph Labs' Physical-AI data pipeline: a head camera plus our OGLO tactile gloves on both hands, with action labels derived from the physical contact signal.… See the full description on the dataset page: https://huggingface.co/datasets/OpenGraphLabs-Research/ego-tactile-manipulation.berkeley_fanuc_manipulation_lerobotThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "fanuc_mate",
"total_episodes": 415,
"total_frames": 62613,
"total_tasks": 32,
"total_videos": 830,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 10,
"splits": {
"train": "0:415"
},
"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/IPEC-COMMUNITY/berkeley_fanuc_manipulation_lerobot.PhysicalAI-Robotics-Manipulation-Kitchen
PhysicalAI Robotics Manipulation in the Kitchen
Dataset Description:
PhysicalAI-Robotics-Manipulation-Kitchen is a dataset of automatic generated motions of robots performing operations such as opening and closing cabinets, drawers, dishwashers and fridges. The dataset was generated in IsaacSim leveraging reasoning algorithms and optimization-based motion planning to find solutions to the tasks automatically [1, 3]. The dataset includes a bimanual manipulator built with… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-Manipulation-Kitchen.rdc-manipulation-datasetsai-election-manipulation-cases
AI, Elections and Agency Transfer Evidence Index
Version 0.4.4 · released 21 August 2026 · research cutoff 12 August 2026
The dataset contains 6 documented-manipulation records, not 1,087 cases. Read the counts in this order:
1,087 relational rows -> 64 catalogue entries -> 10 core records
-> 8 incident-eligible records
-> 6 documented-manipulation records
The other two incident-eligible records are transparent contested-use… See the full description on the dataset page: https://huggingface.co/datasets/apol/ai-election-manipulation-cases.ManipulationVQA-60k
Robo2VLM-Reasoning
Samples from the dataset: Robo2VLM-1, prompting gemini-2.5-pro to generate reasoning traces supporting the correct choice.
Paper: Robo2VLM: Visual Question Answering from Large-Scale In-the-Wild Robot Manipulation Datasets
@misc{chen2025robo2vlmvisualquestionanswering,
title={Robo2VLM: Visual Question Answering from Large-Scale In-the-Wild Robot Manipulation Datasets},
author={Kaiyuan Chen and Shuangyu Xie and Zehan Ma and Ken Goldberg}… See the full description on the dataset page: https://huggingface.co/datasets/keplerccc/ManipulationVQA-60k.berkeley_fanuc_manipulation_rawEB-ManipulationThis repository contains the data of the paper EmbodiedBench: Comprehensive Benchmarking Multi-modal Large Language Models for Vision-Driven Embodied Agents.
📄paper
💻 Github
🏠Website
utokyo_pr2_tabletop_manipulationThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "unknown",
"total_episodes": 240,
"total_frames": 32708,
"total_tasks": 3,
"total_videos": 240,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 5,
"splits": {
"train": "0:240"
},
"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/utokyo_pr2_tabletop_manipulation.berkeley_fanuc_manipulationThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "unknown",
"total_episodes": 415,
"total_frames": 62613,
"total_tasks": 32,
"total_videos": 830,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 10,
"splits": {
"train": "0:415"
},
"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/berkeley_fanuc_manipulation.PhysicalAI-Robotics-Manipulation-ObjectsPhysicalAI-Robotics-Manipulation-Objects is a dataset of automatic generated motions of robots performing operations such as picking and placing objects in a kitchen environment. The dataset was generated in IsaacSim leveraging reasoning algorithms and optimization-based motion planning to find solutions to the tasks automatically [1, 3]. The dataset includes a bimanual manipulator built with Kinova Gen3 arms. The environments are kitchen scenes where the furniture and appliances were… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-Manipulation-Objects.berkeley_fanuc_manipulation_raweveryday-manipulation-3d
everyday-manipulation-3d
Browse the episodes in your browser → · caryx.ai · founders@caryx.ai
Every frame carries metric depth, 6-DoF camera pose, MANO hand pose,
per-hand object contact, hand and object segmentation, and time-aligned
language, so a model can be tested against any one of those channels or all of
them at once without collecting or labelling anything first.
Recorded on iPhone Pro (ARKit LiDAR) as part of Everyday Manipulation 1 (EM1),
annotated by CaryX AI… See the full description on the dataset page: https://huggingface.co/datasets/CaryxAI/everyday-manipulation-3d.conq_hose_manipulation_rawmanipulation_data_3This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "panda",
"total_episodes": 2926,
"total_frames": 1096588,
"total_tasks": 1,
"total_videos": 0,
"total_chunks": 3,
"chunks_size": 1000,
"fps": 20,
"splits": {
"train": "0:2926"
},
"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/lqs2015/manipulation_data_3.isaaclab-manipulation-objectsfactory-manipulation-videos
Factory manipulation videos
Procedural Robotics is open sourcing a small set of our factory data so teams can assess its quality. The videos show workers performing factory tasks.
Contents
Seven continuous takes, 109 minutes in total.
Task
Station
Worker
Duration
File
cardboard manipulation
01
041
23.6 min
cardboard_manipulation_station01_worker041.mp4
cardboard manipulation
04
026
16.5 min
cardboard_manipulation_station04_worker026.mp4
defect… See the full description on the dataset page: https://huggingface.co/datasets/procedural-robotics/factory-manipulation-videos.yam_bimanual_manipulation
YAM Bimanual Manipulation
Real-robot demonstrations for three bimanual manipulation tasks, collected on a
YAM dual-arm platform and stored in
LeRobot v3.0 format.
Each task is a self-contained LeRobot dataset in its own top-level directory:
Directory
Task
Episodes
Frames
Size
blocks_filtered/
Put all blocks into the box.
97
136,489
2.3 GB
dustpan_filtered/
Clean the table using the dust pan.
100
58,018
773 MB
transfer_filtered/
Transfer the egg from the pan into… See the full description on the dataset page: https://huggingface.co/datasets/shailes-h/yam_bimanual_manipulation.1000_Robot_Manipulation_Tasks
1000 Robot Manipulation Tasks
This dataset is currently undergoing a phased upload process. Due to its substantial size, the complete data is being incrementally published to ensure stability and integrity during transmission. We appreciate your patience as we finalize the full availability of the dataset.
This dataset is distributed under the MIT License. This permissive free software license grants users the freedom to use, copy, modify, merge, publish, distribute, sublicense… See the full description on the dataset page: https://huggingface.co/datasets/PitVit/1000_Robot_Manipulation_Tasks.