datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Dexora_Real-World_Dataset
Dexora: Open-Source VLA for High-DoF Bimanual Dexterity
🔥 News & Updates
2025-12-03: Released the full Real-World Dataset (12.2K episodes) on Hugging Face.
2025-12-12: Released the task-level view of the Real-World Dataset (one folder per high-level task) on Hugging Face.
Coming soon: We will open-source the full 100K-episode simulation dataset.
--
📊 Dataset Overview
The Dexora corpus combines high-fidelity real-world teleoperation data with… See the full description on the dataset page: https://huggingface.co/datasets/Dexora/Dexora_Real-World_Dataset.real_world_datareal_world_data
Real-world manipulation demonstrations — common HDF5 interface
This release standardizes 1,000 episodes across five tasks for shared
camera3, robot-state, timestamp, and language loading. There are 200 episodes
per task: 100 clean and 25 each at clutter levels d1, d2, d3, and d4.
Conditions are collection subsets, not predefined train/validation/test splits.
Four older tasks have been upgraded to the HDF5 interface of the newer
three-cups task. The 200 three-cups HDF5 files… See the full description on the dataset page: https://huggingface.co/datasets/mzxuan/real_world_data.arx5_real_world_datasets
arx5_real_world_datasets
This repository contains the real-world robot datasets collected on the ARX-5 platform as presented in the paper Self-Correcting VLA: Online Action Refinement via Sparse World Imagination.
The dataset is part of the SC-VLA framework, which enhances robotic physical grounding through intrinsic self-improvement using Sparse World Imagination (SPI) and Online Action Refinement (OAR).
Code: https://github.com/Kisaragi0/SC-VLA
Paper: Self-Correcting VLA: Online… See the full description on the dataset page: https://huggingface.co/datasets/Kisaragi0/arx5_real_world_datasets.Real-World-SQL-Bench-Databasesvla0-real-world-dataset-v3This 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/mattpidden/vla0-real-world-dataset-v3.real_world_data_smallreal_world_data_annotated
Real-World Data — Camera-3 Annotations
This dataset provides perception annotations for 1,000 real robot demonstration
episodes across five tasks. It is the annotation companion to
mzxuan/real_world_data,
which contains the original recordings.
Download both datasets for visual robot-learning experiments. This repository
does not contain the original RGB, depth, robot state, or action commands. It
contains masks, object identities and roles, 2D/3D boxes, annotation-only HDF5… See the full description on the dataset page: https://huggingface.co/datasets/mzxuan/real_world_data_annotated.vla0-real-world-datasetThis 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/mattpidden/vla0-real-world-dataset.vla0-real-world-dataset-v3_20260622_150943This 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/mattpidden/vla0-real-world-dataset-v3_20260622_150943.realworld_dataTunisian-Mediterranean-Coastal-Marine-Litter-Trash-Dataset-Drone-Based-Real-World-Images50hours_Malay_Real-world_Colloquial_Conversation_and_Monologue_Speech_Dataset
BabelSpeech: 50 Hours of Real-World Colloquial Malay ASR Speech Data
This dataset contains 50 hours of high-quality Malay colloquial ASR speech data, reflecting realistic code-switching between Malay and English, as commonly used in everyday communication in Malaysia.
Overview
Content: 50 hours of real-world colloquial Malay speech suitable for ASR fine-tuning and benchmarking.
Metadata: Stored in a separate JSON file, including audio path, duration, text, confidence… See the full description on the dataset page: https://huggingface.co/datasets/BabelSpeech/50hours_Malay_Real-world_Colloquial_Conversation_and_Monologue_Speech_Dataset.Real-World-Dataset
Latest Updates
[2025-10-23] v0.3 Dataset Released. 304 episodes across five tasks. This dataset was collected using two separate evaluation robot arms (one in Beijing and one in Shanghai). This can be understood as data from systems that have slight differences in their kinematic or sensor calibration parameters.
[2025-10-21] We have released the real-bot mcap data packaging tool (supporting both Arrow and LMDB formats). This tool can package mcap files recorded by the Data… See the full description on the dataset page: https://huggingface.co/datasets/HorizonRobotics/Real-World-Dataset.realworld_dataset_backgroundThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "bifranka",
"total_episodes": 1,
"total_frames": 54,
"total_tasks": 1,
"total_videos": 1,
"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/younghyopark/realworld_dataset_background.realworld_dataset_0708_2This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "bifranka",
"total_episodes": 1,
"total_frames": 645,
"total_tasks": 1,
"total_videos": 1,
"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/younghyopark/realworld_dataset_0708_2.vla0-real-world-dataset-v2_20260622_150036This 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/mattpidden/vla0-real-world-dataset-v2_20260622_150036.Reindex-Then-Adapt-Real-World-Dataabhijitdahatonde_real-world-smartphones-dataset
Real World Smartphone's Dataset
Worlds Smartphones: A Comprehensive Dataset for Cutting-Edge Analysis
Dataset Info
Source: Kaggle
Original Size: 0.02 MB
Kaggle Downloads: 5,744
Files: 1
Files
smartphones.csv
Mirrored from Kaggle
realworld_dataset_0708This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "bifranka",
"total_episodes": 1,
"total_frames": 405,
"total_tasks": 1,
"total_videos": 1,
"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/younghyopark/realworld_dataset_0708.realworld_datavla0-real-world-dataset_20260610_110151real-world-medical-mistakes-dataset
Real-World Medical Mistakes Dataset
A curated dataset of 100 de-identified clinical reports from Internal Medicine and Emergency Departments, each containing a physician-inserted realistic medical error. Designed for training and evaluating AI systems that detect critical patient safety errors in clinical documentation.
Dataset Description
Overview
This dataset was created as part of the Clinipal project — an AI-powered clinical error detection system. Three… See the full description on the dataset page: https://huggingface.co/datasets/Vrda/real-world-medical-mistakes-dataset.auvire-real-world-data
Citation
This model is from the paper "AuViRe: Audio-visual Speech Representation Reconstruction for Deepfake Temporal Localization" available at https://huggingface.co/papers/2511.18993
Master_Bedroom_Real_World_Residential_Scene_DatasetRealWorldDatasetSplit_final_setreal-world-datareal-world-data-v2