datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
EgoDex-LeRobot-v3.0EgoDex dataset from https://github.com/apple/ml-egodex, converted to LeRobot datasets v3.0 format.
This repo is structured with train / test directories at the top level,
under each of which are multiple subdirectories, each corresponding to one directory in the original EgoDex data format.
Each of these represents one LeRobot dataset, so to load the full EgoDex training set you have to load each subdirectory
as its own LeRobotDataset and then combine them (e.g. w/ PyTorch's ConcatDataset), or… See the full description on the dataset page: https://huggingface.co/datasets/griffinlabs/EgoDex-LeRobot-v3.0.Galaxea-Open-World-Dataset-LeRobot-v3.0Galaxea Open-World Dataset taken from OpenGalaxea/Galaxea-Open-World-Dataset,
converted to LeRobot Datasets v3.0 format using lerobot.datasets.v30.convert_dataset_v21_to_v30.
Missing subsets
The subset Boil_The_Water_20250714_006 is missing due to the original files having some episodes at 62 fps,
which causes the conversion script to crash with an error.
The subset Put_The_Items_Into_The_Storage_Box_20250929_002_007 is missing due to it having 7 DoF arms rather than 6 DoF.… See the full description on the dataset page: https://huggingface.co/datasets/griffinlabs/Galaxea-Open-World-Dataset-LeRobot-v3.0.InternData-A1-LeRobot-v3.0-by-embodimentInternData-A1 dataset taken from InternRobotics/InternData-A1,
with the tarballs extracted and directory structure "transposed" so that the top-level subdirectories are the four embodiments.
Two franka dirs
For the franka embodiment, there are two different feature spaces, so we split it into the franka-1 and franka-2 directories.
The feature spaces differ in image shape and gripper value range.
Minor fixes
Some subsets such as… See the full description on the dataset page: https://huggingface.co/datasets/griffinlabs/InternData-A1-LeRobot-v3.0-by-embodiment.anything-v3.0-glazed
Dataset Card for Anything v3.0 Glazed Samples
Dataset Description
Dataset Summary
This dataset contains image samples originally generated by Linaqruf/anything-v3.0
and subsequently processed by Glaze tool.
Supported Tasks and Leaderboards
[More Information Needed]
Languages
[More Information Needed]
Dataset Structure
Data Instances
[More Information Needed]
Data Fields
[More Information Needed]… See the full description on the dataset page: https://huggingface.co/datasets/hanamizuki-ai/anything-v3.0-glazed.RoboTwin-LeRobot-v3.0sphere_cohere_embed-english-v3.0refinedweb-embed-english-v3.0maestro-v3.0.0-preprocessedmaestro-v3.0.0
Maestro Dataset v3.0.0
Mirror copy as of 08/03/2025
MAESTRO (MIDI and Audio Edited for Synchronous TRacks and Organization) is a dataset composed of about 200 hours of virtuosic piano performances captured with fine alignment (~3 ms) between note labels and audio waveforms.
Source download Maestro Dataset v3.0.0
Citation
@inproceedings{
hawthorne2018enabling,
title={Enabling Factorized Piano Music Modeling and Generation with the {MAESTRO}… See the full description on the dataset page: https://huggingface.co/datasets/projectlosangeles/maestro-v3.0.0.cleaning_table_hdf5-v3.0maestro-v3.0.0MAESTRO-v3.0-FLACrobocasa365-lerobot-v3.0
RoboCasa365 LeRobot v3.0
This dataset packages the RoboCasa365 pretraining demonstrations used by EasyWAM in LeRobot v3.0 format. It includes the atomic and composite pretraining corpora with synchronized multi-view RGB video, robot state, actions, rewards, termination flags, and task annotations.
Dataset Summary
Subset
Episodes
Frames
Tasks
Pretrain Atomic
7,356
1,495,313
617
Pretrain Composite
24,687
27,610,913
4,428
Total
32,043
29,106,226
5,045… See the full description on the dataset page: https://huggingface.co/datasets/OpenMOSS-Team/robocasa365-lerobot-v3.0.robotwin2.0-lerobot-v3.0
RoboTwin 2.0 LeRobot v3.0
This dataset packages RoboTwin 2.0 bimanual-manipulation demonstrations in LeRobot v3.0 format for EasyWAM training. It contains synchronized high-view and dual-wrist RGB video, bimanual joint state, actions, and natural-language task metadata.
Dataset Summary
Episodes
Frames
Videos
Recording rate
27,500
6,075,103
82,500
50 FPS
Structure
robotwin2.0-lerobot-v3.0/
├── data/ # frame-level… See the full description on the dataset page: https://huggingface.co/datasets/OpenMOSS-Team/robotwin2.0-lerobot-v3.0.wikipedia_0715_clean_cohere_embed-english-v3.0maestro-v3.0.0-preprocessed-with-pedal-2lm-eval-results-shyamieee-Padma-SLM-7b-v3.0-private
Dataset Card for Evaluation run of shyamieee/Padma-SLM-7b-v3.0
Dataset automatically created during the evaluation run of model shyamieee/Padma-SLM-7b-v3.0
The dataset is composed of 62 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 2 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An… See the full description on the dataset page: https://huggingface.co/datasets/nyu-dice-lab/lm-eval-results-shyamieee-Padma-SLM-7b-v3.0-private.robodojo-lerobot-v3.0
RoboDojo LeRobot v3.0
This dataset packages RoboDojo robot-manipulation demonstrations in LeRobot v3.0 format for EasyWAM training. It contains synchronized high-view and dual-wrist RGB video, bimanual joint state, actions, and natural-language task metadata.
Dataset Summary
Episodes
Frames
Tasks
Recording rate
3,500
1,856,102
35
25 FPS
Structure
robodojo-lerobot-v3.0/
├── data/ # frame-level Parquet files
├── meta/… See the full description on the dataset page: https://huggingface.co/datasets/OpenMOSS-Team/robodojo-lerobot-v3.0.libero_10_lerobot_v3.0This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "panda",
"total_episodes": 379,
"total_frames": 101469,
"total_tasks": 10,
"chunks_size": 1000,
"fps": 10,
"splits": {
"train": "0:379"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4"… See the full description on the dataset page: https://huggingface.co/datasets/zeromidnight/libero_10_lerobot_v3.0.libero-lerobot-v3.0
LIBERO LeRobot v3.0
This dataset packages the four standard LIBERO demonstration suites in LeRobot v3.0 format for robot-policy and world-action-model training. It contains synchronized RGB observations, robot state, actions, timestamps, task indices, and natural-language task metadata.
Dataset Summary
Suite
Episodes
Frames
Tasks
Videos
LIBERO-Spatial
434
53,229
10
868
LIBERO-Object
457
67,309
10
914
LIBERO-Goal
433
52,895
10
866
LIBERO-10
388
104… See the full description on the dataset page: https://huggingface.co/datasets/OpenMOSS-Team/libero-lerobot-v3.0.libero_object_no_noops_1.0.0_lerobot_v3.0
libero_object_no_noops_1.0.0_lerobot
Dataset Description
This dataset is the LIBERO-Object subset of the LIBERO benchmark, converted to LeRobot v3.0 format.
10 object generalization tasks. All tasks share the same spatial layout but involve different objects, testing the agent's ability to generalize to novel object instances.
Note: This dataset has been filtered to remove no-op frames (idle frames where the robot does not execute meaningful actions). This results in… See the full description on the dataset page: https://huggingface.co/datasets/hello3x3/libero_object_no_noops_1.0.0_lerobot_v3.0.miracl-th-corpus-multilingual-v3.0details_migtissera__Synthia-7B-v3.0
Dataset Card for Evaluation run of migtissera/Synthia-7B-v3.0
Dataset automatically created during the evaluation run of model migtissera/Synthia-7B-v3.0 on the Open LLM Leaderboard.
The dataset is composed of 63 configuration, each one coresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard-old/details_migtissera__Synthia-7B-v3.0.stackexchange_cohere_embed-english-v3.0Stuffed_Animal_V3.0
Stuffed_Animal_V3.0
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.
maestro-v3.0.0-preprocessed-with-pedallibero_90_no_noops_lerobot_v3.0
libero_90_no_noops_lerobot
Dataset Description
This dataset is the LIBERO-90 subset of the LIBERO benchmark, converted to LeRobot v3.0 format.
A large-scale set of 73 diverse manipulation tasks intended for pre-training. Covers a broad range of manipulation skills including pick-and-place, drawer/cabinet interaction, stacking, and tool use.
Note: This dataset has been filtered to remove no-op frames (idle frames where the robot does not execute meaningful actions). This… See the full description on the dataset page: https://huggingface.co/datasets/hello3x3/libero_90_no_noops_lerobot_v3.0.robodojo_lerobot_v3.0libero_10_no_noops_1.0.0_lerobot_v3.0
libero_10_no_noops_1.0.0_lerobot
Dataset Description
This dataset is the LIBERO-Long subset of the LIBERO benchmark, converted to LeRobot v3.0 format.
10 long-horizon multi-step manipulation tasks. Each task requires executing a sequence of subtasks in order, testing the agent's ability to plan and compose multiple skills.
Note: This dataset has been filtered to remove no-op frames (idle frames where the robot does not execute meaningful actions). This results in more… See the full description on the dataset page: https://huggingface.co/datasets/hello3x3/libero_10_no_noops_1.0.0_lerobot_v3.0.Hercules-v3.0
Hercules-v3.0
Dataset Name: Hercules-v3.0
Version: 3.0
Release Date: 2024-2-14
Number of Examples: 1,637,895
Domains: Math, Science, Biology, Physics, Instruction Following, Conversation, Computer Science, Roleplay, and more
Languages: Mostly English, but others can be detected.
Task Types: Question Answering, Conversational Modeling, Instruction Following, Code Generation, Roleplay
Data Source Description
Hercules-v3.0 is an extensive and diverse dataset that… See the full description on the dataset page: https://huggingface.co/datasets/Locutusque/Hercules-v3.0.