datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
droid_v3This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "Franka",
"total_episodes": 95658,
"total_frames": 27630375,
"total_tasks": 49630,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 15,
"splits": {
"train": "0:95658"
},
"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/DAVIAN-Robotics/droid_v3.speech-commands-v0.02
Speech Commands Dataset v0.02
This is a re-hosted copy of the Google Speech Commands v0.02 dataset in Parquet format for compatibility with the Hugging Face Dataset Viewer.
⚠️ Credits
This dataset was created by Pete Warden / Google. All credit goes to the original authors and the crowdsourcing contributors.
Original source: http://download.tensorflow.org/data/speech_commands_v0.02.tar.gz
Paper: Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition… See the full description on the dataset page: https://huggingface.co/datasets/pollen-robotics/speech-commands-v0.02.robolab-init-positionsfactory-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.CausalVerse_Video_Robotics_Kitchen
CausalVerse Video Dataset
Available splits: robotics_kitchen
Each record contains the following columns:
videos, metavalue, npz_data
robotics_kitchen
Examples: 7500
Columns: videos, metavalue, npz_data
qualia-dataset-real
Qualia Real-Robot Datasets
Release
release-2026-07-19.3 (changelog)
Domain
Real-robot recordings
Format
LeRobot v2.1
Access
Public
Contact
Start a discussion
Public access makes the recordings inspectable and reproducible; usage rights still follow each
source dataset. See License and provenance.
Dataset description
This repository collects Qualia's real-robot datasets in one place. Each dataset contains one
instruction and records the… See the full description on the dataset page: https://huggingface.co/datasets/qualia-robotics/qualia-dataset-real.CausalVerse_Video_Robotics_Living
CausalVerse Video Dataset
Available splits: robotics_living
Each record contains the following columns:
videos, metavalue, npz_data
robotics_living
Examples: 4800
Columns: videos, metavalue, npz_data
CausalVerse_Video_Robotics_Mobile
CausalVerse Video Dataset
Available splits: robotics_mobile
Each record contains the following columns:
videos, metavalue, npz_data
robotics_mobile
Examples: 2001
Columns: videos, metavalue, npz_data
qualia-dataset-sim
Qualia Simulation Datasets
Release
release-2026-07-19.3 (changelog)
Domain
Robot simulation
Format
LeRobot v2.1
Access
Public
Contact
Start a discussion
This collection is public. Provenance and source licensing are recorded in each dataset's
dataset.yaml.
Dataset description
This repository collects Qualia's simulated robot datasets in one place. Each dataset contains one
language instruction and records the source revision used to… See the full description on the dataset page: https://huggingface.co/datasets/qualia-robotics/qualia-dataset-sim.CausalVerse_Video_Robotics_Study
CausalVerse Video Dataset
Available splits: robotics_study
Each record contains the following columns:
videos, metavalue, npz_data
robotics_study
Examples: 2700
Columns: videos, metavalue, npz_data
robotics-quality-leaderboard
Robotics Dataset Quality Leaderboard
What This Is
This repository hosts an automatically-updated quality leaderboard for robotics
imitation-learning datasets on HuggingFace. Each dataset is scored by the
HaptalAI quality scorer, an
open-source tool that streams a sample of episodes from a dataset, detects the
available sensor schema, runs a suite of failure-detection checks, and computes
a single 0–100 quality score. The leaderboard is intended as a first-pass… See the full description on the dataset page: https://huggingface.co/datasets/HaptalAI/robotics-quality-leaderboard.LR_Robotics_realsmolvla-eval-devLR_RoboticsCausalVerse_Video_Robotics_General
CausalVerse Video Dataset
Available splits: robotics_general
Each record contains the following columns:
videos, metavalue, npz_data
robotics_general
Examples: 4500
Columns: videos, metavalue, npz_data
LR_Robotics_real_diagonal_v8YOLOv8-Object-Detection-02-Dataset
Soup Can Object Detection Dataset Sample
Duality.ai just released a 1000 image dataset used to train a YOLOv8 model for object detection -- and it's 100% free!
Just create an EDU account here.
This HuggingFace dataset is a 20 image and label sample, but you can get the rest at no cost by creating a FalconCloud account. Once you verify your email, the link will redirect you to the dataset page.
Dataset Overview
This dataset consists of high-quality images of soup cans… See the full description on the dataset page: https://huggingface.co/datasets/duality-robotics/YOLOv8-Object-Detection-02-Dataset.semsim_1k
semsim_1k
A 1,427-image robotics corpus for specification-grounded runtime safety
monitoring, curated from five public sources and unified under one schema.
This is the annotation and human-review corpus.
Each image is a frame a mobile robot could plausibly have captured, kept with
enough context that a human can judge whether a described hazard is really
visible in it.
Loading — sources are splits, not configs
from datasets import load_dataset
gt =… See the full description on the dataset page: https://huggingface.co/datasets/cairo-robotics/semsim_1k.YOLOv8-Multiclass-Object-Detection-Dataset
DATASET SAMPLE
Duality.ai just released a 1000 image dataset used to train a YOLOv8 model in multiclass object detection -- and it's 100% free!
Just create an EDU account here.
This HuggingFace dataset is a 20 image and label sample, but you can get the rest at no cost by creating a FalconCloud account. Once you verify your email, the link will redirect you to the dataset page.
What makes this dataset unique, useful, and capable of bridging the Sim2Real gap?
The digital twins are… See the full description on the dataset page: https://huggingface.co/datasets/duality-robotics/YOLOv8-Multiclass-Object-Detection-Dataset.social-robotics-saf
Social Robotics — Social-Affective Filter (SAF), combined metadata
Dehydrated social-signal metadata from 200 egocentric Ego4D clips — the combined output of the Social-Affective Filter (SAF) layers in one wide table, so robots can learn to read human reactions. No raw video pixels and no audio. Each row is one source video keyed by video_id; rehydrate against your own legally-obtained Ego4D copies (below).
Each layer is also published standalone with a detailed card:
attention… See the full description on the dataset page: https://huggingface.co/datasets/louisye/social-robotics-saf.robotics-embodied-ai-physical-control-2026
🤖 Robotics, Embodied AI & Physical World Control Dataset (2026 Edition)
A structured research dataset featuring 2,123 domain-verified research papers and official code repositories focused on Vision-Language-Action Models (VLA), Humanoid Robotics, Quadruped Locomotion, Diffusion Policies, Sim-to-Real Transfer, and Physics Simulation Environments (Isaac Sim, MuJoCo, Genesis).
Built with Universal Scientific Engine V16.1 Gold, providing 43 schema attributes with verified… See the full description on the dataset page: https://huggingface.co/datasets/beatsprom/robotics-embodied-ai-physical-control-2026.flock-robotics-vla-training-v2
FLock Robotics VLA Training Dataset v2
Expert demonstrations for the FLock Robotics VLA competition task. All trajectories are successful.
Quick start
from datasets import load_dataset
ds = load_dataset("random-sequence/flock-robotics-vla-training-v2")
print(ds["train"][0])
# Keys: episode_index, step_index, task, difficulty, instruction,
# image (PIL), action [7], proprio [25], reward, done
Dataset statistics
Task
Episodes
Difficulty… See the full description on the dataset page: https://huggingface.co/datasets/random-sequence/flock-robotics-vla-training-v2.LR_Robotics_real_v2sustainability-roboticsYOLOv8-Multi-Instance-Object-Detection-Dataset
Multi Instance Object Detection Dataset Sample
Duality.ai just released a 1000 image dataset used to train a YOLOv8 model for object detection -- and it's 100% free!
Just create an EDU account here.
This HuggingFace dataset is a 20 image and label sample, but you can get the rest at no cost by creating a FalconCloud account. Once you verify your email, the link will redirect you to the dataset page.
Dataset Overview
This dataset consists of high-quality images of soup… See the full description on the dataset page: https://huggingface.co/datasets/duality-robotics/YOLOv8-Multi-Instance-Object-Detection-Dataset.HumanDrive-CARLA
G29-HumanDrive-CARLA-5Runs
Dataset Description
This dataset contains 70,308 samples collected in CARLA using a Logitech G29 steering wheel.
It includes one representative run from each town:
runs_01 → Town01
runs_02 → Town03
runs_03 → Town04
runs_04 → Town05
runs_05 → Town07
Each sample contains:
RGB image embedded in JPEG
semantic segmentation image embedded in PNG
steering command
throttle command
brake command
DAgger flag
town label
run identifier
frame identifier… See the full description on the dataset page: https://huggingface.co/datasets/roboticslaburjc/HumanDrive-CARLA.social-robotics-attention
Social Robotics: Attention / Engagement (03a)
Does the bystander look at the camera-wearer? Per-bystander gaze + head-pose engagement around each task.
One layer of the Social-Affective Filter (SAF) — dehydrated social-signal metadata extracted from egocentric (first-person) video so robots can learn to read human reactions. No raw pixels and no audio. Each row is one source video, keyed by video_id; rehydrate against your own legally-obtained Ego4D copies (below).
Rows: 828… See the full description on the dataset page: https://huggingface.co/datasets/louisye/social-robotics-attention.robotics-failure-benchmark
Haptal Robotics Failure Benchmark v1.0
The first public benchmark for robot training data annotation quality and failure detection in manipulation episodes.
What this is
A held-out test set of 600 robot episodes across 6 failure classes generated from real LeRobot trajectories with physics-based failure injection. The test set is fixed. Anyone can evaluate their annotation pipeline against it and get a comparable score.
Why it exists
No standardized… See the full description on the dataset page: https://huggingface.co/datasets/HaptalAI/robotics-failure-benchmark.social-robotics-affirmation-gesture
Social Robotics: Affirmation Gesture (03e)
Did the bystander nod (affirm) or shake their head (negate)?
One layer of the Social-Affective Filter (SAF) — dehydrated social-signal metadata extracted from egocentric (first-person) video so robots can learn to read human reactions. No raw pixels and no audio. Each row is one source video, keyed by video_id; rehydrate against your own legally-obtained Ego4D copies (below).
Rows: 320 — videos in the evaluation slice for which this… See the full description on the dataset page: https://huggingface.co/datasets/louisye/social-robotics-affirmation-gesture.social-robotics-proxemic-kinematics
Social Robotics: Proxemic Kinematics (03d)
Did the bystander move toward (approach) or away from (retreat) the camera-wearer?
One layer of the Social-Affective Filter (SAF) — dehydrated social-signal metadata extracted from egocentric (first-person) video so robots can learn to read human reactions. No raw pixels and no audio. Each row is one source video, keyed by video_id; rehydrate against your own legally-obtained Ego4D copies (below).
Rows: 862 — videos in the evaluation… See the full description on the dataset page: https://huggingface.co/datasets/louisye/social-robotics-proxemic-kinematics.