datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
grasp-datasetUMI-Gripper-Grasping
UMI-Gripper-Grasping
This repository contains the datasets used for the project UMI-Gripper-Grasping, which investigates contact-aware grasp regulation for deformable object manipulation using GelSight tactile sensing, force/torque sensing, and leader–follower teleoperation.
The repository contains two sub-datasets:
Force-State Training Data – labeled tactile and force recordings used to train the three-class XGBoost force-state classifier.
LeRobot-UMI-Integration –… See the full description on the dataset page: https://huggingface.co/datasets/yuezh23/UMI-Gripper-Grasping.Grasp-Any-Region-Dataset
Grasp Any Region Dataset
This repository contains the training dataset for the paper: Grasp Any Region: Towards Precise, Contextual Pixel Understanding for Multimodal LLMs.
Code: https://github.com/Haochen-Wang409/Grasp-Any-Region
About the Dataset
The Grasp Any Region (GAR) dataset is designed to empower Multimodal Large Language Models (MLLMs) with comprehensive region-level visual understanding. While MLLMs excel at holistic understanding, they often struggle with… See the full description on the dataset page: https://huggingface.co/datasets/HaochenWang/Grasp-Any-Region-Dataset.grare-graspnetGrasp-Anything-6D
Grasp-Anything-6D Dataset
We introduce the Grasp-Anything-6D dataset. Building upon our previous Grasp-Anything dataset, Grasp-Anything-6D is a large-scale dataset for the task of language-driven 6-DoF grasp detection.
Our ECCV2024 paper can be accessed at arXiv.
Dataset Structure:
The dataset is structured as follows:
pc: contains .npy files of point cloud scenes. Each scene consists of 8192 points.
pc_mask: contains .npy files of 3D masks for objects in point cloud… See the full description on the dataset page: https://huggingface.co/datasets/airvlab/Grasp-Anything-6D.graspclutter6d
Dataset Card for graspclutter6d
GraspClutter6D is a large-scale real-world dataset for robust perception and grasping in cluttered scenes. This FiftyOne dataset contains a curated subset of 99 scenes from the full GraspClutter6D dataset, optimized for visualization and exploration.
This is a FiftyOne dataset with 111 samples.
Installation
If you haven't already, install FiftyOne:
pip install -U fiftyone
Usage
import fiftyone as fo
from huggingface_hub… See the full description on the dataset page: https://huggingface.co/datasets/Voxel51/graspclutter6d.GraspClutter6D
GraspClutter6D Dataset
GraspClutter6D: A Large-scale Real-world Dataset for Robust Perception and Grasping in Cluttered Scenes
Seunghyeok Back, Joosoon Lee, Kangmin Kim, Heeseon Rho, Geonhyup Lee, Raeyoung Kang, Sangbeom Lee, Sangjun Noh, Youngjin Lee, Taeyeop Lee, Kyoobin Lee
Accepted at IEEE Robotics and Automation Letters (RA-L)
[ArXiv] [Project Website] [Video] [Dataset Format] [GitHub]
To download the dataset, run the following commands:mkdir GraspClutter6D && cd GraspClutter6D… See the full description on the dataset page: https://huggingface.co/datasets/GraspClutter6D/GraspClutter6D.Multi-GraspSet
Multi-GraspLLM: A Multimodal LLM for Multi-Hand Semantic-Guided Grasp Generation
Project Page | arXiv
Updates
2025.3: We add the Jaco hand data which is our final version.
2025.1: We add the Barrett hand data.
2024.12: Released the Multi-GraspSet dataset with meshes of objects.
2024.12: Released the Multi-GraspSet dataset with contact annotations.
Overview
We introduce Multi-GraspSet, the first large-scale multi-hand grasp dataset enriched with automatic… See the full description on the dataset page: https://huggingface.co/datasets/LHS-LHS/Multi-GraspSet.GraspBox-v0GraspNet-1Billion
GraspNet-1Billion (Unofficial Mirror)
⚠️ DISCLAIMER / 声明:
This is an UNOFFICIAL mirror/upload of the GraspNet-1Billion dataset. This repository is maintained solely for ease of access and research reproducibility.
I do not own the rights to this data. All rights belong to the original GraspNet team, MVIG, and SJTU.
🔗 Official Sources / 官方源
Please always refer to the official website for the latest updates, and documentation:
Official Website:… See the full description on the dataset page: https://huggingface.co/datasets/DravenALG/GraspNet-1Billion.Grasp-Anything
Grasp-Anything Dataset
Welcome to the Grasp-Anything dataset!
We are excited to release the Grasp-Anything dataset, comprising 1 million samples and featuring 3 million objects. This dataset aims to facilitate research and development in robotic grasping tasks.
Please fill out the download form and agree to our license prior to downloading the dataset.
The Grasp-Anything++ dataset with grasping prompts is available here.
Dataset Structure:
The dataset is structured as… See the full description on the dataset page: https://huggingface.co/datasets/airvlab/Grasp-Anything.PhysicalAI-Robotics-GraspGen
GraspGen: Scaling Sim2Real Grasping
GraspGen is a large-scale simulated grasp dataset for multiple robot embodiments and grippers.
We release over 57 million grasps, computed for a subset of 8515 objects from the Objaverse XL (LVIS) dataset. These grasps are specific to three grippers: Franka Panda, the Robotiq-2f-140 industrial gripper, and a single-contact suction gripper (30mm radius).
Dataset Format
The dataset is released in the WebDataset format. The… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/PhysicalAI-Robotics-GraspGen.GRASPNET_2grasp-dataset-curated
🖐️ Synthetic Grasp Dataset (Objaverse-LVIS Curated)
This dataset contains high-quality synthetic grasp data generated for robotic manipulation research.
It focuses on the fusion of vision and tactile sensing by providing visibility and occlusion analysis for each contact point.
📊 Dataset Statistics
Number of objects: 10
Source: Curated objects from Objaverse-LVIS (categories: cup, bottle, hammer, screwdriver, wrench)
Grasp Strategies: front_back
Camera Resolution:… See the full description on the dataset page: https://huggingface.co/datasets/jack635/grasp-dataset-curated.fod_shape_grasp_2cam_20260920This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 25,
"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/LTH4663/fod_shape_grasp_2cam_20260920.tommyzihao_kitchenware_grasp
TommyZihao Kitchenware Grasp
LeRobot dataset uploaded from the local plate collection.
Codebase version: v3.0
Robot type: so_follower
FPS: 30
Episodes: 100
Frames: 32670
Tasks: 1
This dataset uses the LeRobot v3 layout with meta, data, and videos directories.
reach_grasp_dataset
Reach&Grasp Dataset
A multimodal dataset of the whole upper-limb during simple and complex movements.
Source
This dataset is redistributed from the IIT Dataverse under the CC BY 4.0 license.
Citation
If you use this dataset, please cite the original authors:
Di Domenico, Dario; Forsiuk, Inna; Müller-Cleve, Simon; Tanzarella, Simone; Garro, Florencia; Marinelli, Andrea; Canepa, Michele; Laffranchi, Matteo; Chiappalone, Michela; Bartolozzi, Chiara; De… See the full description on the dataset page: https://huggingface.co/datasets/zhuyu-cs/reach_grasp_dataset.cluttered_grasp
Cluttered Grasp (UR10 + Robotiq 2F-140 + tactile + depth)
Real-robot teleop demonstrations of cluttered grasping. Episodes are gated on
gripper_tcp height in base_link (record while low, pause while high), time-synced
to the RealSense RGB master @ 30 Hz, then converted to
LeRobot format.
Robot
UR10 + Robotiq 2F-140 (ur10_robotiq_2f140)
FPS
30 (RGB master timeline)
Episodes / frames
See meta/info.json (total_episodes, total_frames)
Codebase
LeRobot dataset v3.0… See the full description on the dataset page: https://huggingface.co/datasets/benudavis/cluttered_grasp.G1_Dex3_GraspSquare_DatasetThis dataset was created using LeRobot.
Due to the inability to precisely describe spatial positions, adjust the scene to closely match the first frame of the dataset after installing the hardware as specified in Part 5 of AVP Teleoperation Documentation.
Data collection is not completed in a single session, and variations between data entries exist. Ensure these variations are accounted for during model training.
Dataset Structure
meta/info.json:
{
"codebase_version":… See the full description on the dataset page: https://huggingface.co/datasets/unitreerobotics/G1_Dex3_GraspSquare_Dataset.so100_grasp_pink_blockThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "so100",
"total_episodes": 50,
"total_frames": 43255,
"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/speedyyoshi/so100_grasp_pink_block.G1_Brainco_GraspRubiksCube_DatasetThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "Unitree_G1_Brainco",
"total_episodes": 197,
"total_frames": 220788,
"total_tasks": 1,
"total_videos": 788,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:197"
},
"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/unitreerobotics/G1_Brainco_GraspRubiksCube_Dataset.G1_Brainco_GraspOreo_DatasetThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "Unitree_G1_Brainco",
"total_episodes": 201,
"total_frames": 234959,
"total_tasks": 1,
"total_videos": 804,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:201"
},
"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/unitreerobotics/G1_Brainco_GraspOreo_Dataset.hsl_leap_grasp_rubber_duck_rrt_filtered_v1act-grasp-stack-transferso101-chess-grasp-approach
SO-101 chess research captures
Versioned capture artifacts for the SO-101 chess project.
The current encoder-grasp schema contains six-joint actions, six-joint encoder observations, synchronized 1280x720 wrist and nearest-overhead video, timestamps, frame/episode indices, and task labels. Capture bundles include accepted and negative grasp approaches, legacy placement returns, and the newer positive/negative centered-placement collections. Centered placement journals add… See the full description on the dataset page: https://huggingface.co/datasets/Stefanobraghetto/so101-chess-grasp-approach.grasp_bag_strap_dr_v1
grasp_bag_strap_dr_v1
Bimanual grasp-the-strap-and-lift on a deformable non-woven bag, collected in simulation
(Newton SolverStyle3D cloth + SolverMuJoCo robot co-simulation, Omniverse Kit rendering).
Two randomisation axes vary per episode: the bag appearance (10 texture variants of one
mother asset) and the strap stretch/shear stiffness. Everything else is held fixed.
Contents
Episodes / frames
200 / 83,566
FPS
25
Cameras
static_cam… See the full description on the dataset page: https://huggingface.co/datasets/deformable-bench/grasp_bag_strap_dr_v1.fast-food-floor-waste-grasping-training-set-next-pack-9f7b7681-1106dcde
Fast-Food Cleaning Robot — Floor Mess Dataset
Training dataset for a cleaning robot operating in fast-food-style food-service spaces (break areas / dining). Scenes are staged in break-area environments cluttered with food-service furnishings and food items (pizza, grocery food, cups, spoons) so the robot learns to perceive and act on mess. Covers detection, grasping, navigation, obstacle avoidance and pick-and-place. Renders are 1024x1024 with RGB plus albedo, metric depth and… See the full description on the dataset page: https://huggingface.co/datasets/physicl-community/fast-food-floor-waste-grasping-training-set-next-pack-9f7b7681-1106dcde.Franka_GraspNet_Testact-grasp-stack-transfer-noisyteam-7-right-arm-grasp-tapeThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "so100",
"total_episodes": 100,
"total_frames": 59711,
"total_tasks": 1,
"total_videos": 200,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"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/roboticshack/team-7-right-arm-grasp-tape.