datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
droidThis dataset was created using LeRobot.
DROID: A Large-Scale In-the-Wild Robot Manipulation Dataset
One of the biggest open-source dataset for robotics with 27.044,326 frames, 92,223 episodes, 31,308 unique task description in natural language.
Ported from Tensorflow Dataset format (2TB) to LeRobotDataset format (400GB) with the help from IPEC-COMMUNITY.
Visualization: LeRobot
Homepage: Droid
Paper: Arxiv
License: apache-2.0
Dataset Structure
meta/info.json:
{… See the full description on the dataset page: https://huggingface.co/datasets/cadene/droid.DreamZero-DROID-DataCosmos3-DROID
DROID: Distributed Robot Interaction Dataset
Dataset Summary
DROID (Distributed Robot Interaction Dataset) is a large-scale "in-the-wild" robot manipulation dataset containing 76K teleoperated demonstration trajectories — approximately 350 hours of interaction data — collected across 564 unique scenes, 86 tasks, and 52 buildings over the course of 12 months. The data was collected by 50 data collectors at 18 labs across 13 institutions in North America, Asia, and… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Cosmos3-DROID.droid_1.0.1This 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": 500,
"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/lerobot/droid_1.0.1.droid_failure
DROID Failure Episodes
Failure episodes in DROID-COMMUNITY. Episodes with inconsistent data were discarded, leaving ~91% of all failure episodes.
Ported from raw 1.0.1 data at full resolution to LeRobotDataset v3.0 format (0.25 TiB | 1.5k inodes) with extra annotation from KarlP/droid.
Your browser does not support the video tag.
For the successful episodes, see jnogga/droid_success.
Updates
22.02.2026: Removed episodes with unexpected camera… See the full description on the dataset page: https://huggingface.co/datasets/jnogga/droid_failure.droid_360p
DROID 1.0.1 — 640×360, packed (LeRobot v3.0)
A 640×360 re-encode of the DROID robot-manipulation dataset in
LeRobot v3.0 format, with videos packed into ~500 MB files like the original
lerobot/droid_1.0.1. It reuses droid_1.0.1's exact
tabular data (states / actions / camera extrinsics / task instructions) and replaces the low-resolution
(320×180) video with 640×360 frames decoded from the original DROID raw MP4s (left-eye mono per camera).
What's different from… See the full description on the dataset page: https://huggingface.co/datasets/XinxuanLuM/droid_360p.droid_100This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "unknown",
"total_episodes": 100,
"total_frames": 32212,
"total_tasks": 47,
"total_videos": 300,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 15,
"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/lerobot/droid_100.droid_1.0.1_testThis 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/aractingi/droid_1.0.1_test.MolmoAct2-DROID-Dataset
MolmoAct2-DROID Dataset
This dataset was created using LeRobot.
Language Annotations
This dataset includes annotated language instructions in meta/tasks_annotated.parquet. The file is indexed by episode_index and has a task column containing our per-episode annotated instruction.
The standard LeRobot loader resolves a frame's language instruction through task_index: each data row stores a task_index, which is looked up in meta/tasks.parquet. When you use these… See the full description on the dataset page: https://huggingface.co/datasets/allenai/MolmoAct2-DROID-Dataset.droid_1.0.1_v30This 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": 500,
"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/muacha/droid_1.0.1_v30.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.droid_1.0.1_v30This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "Franka",
"total_episodes": 95584,
"total_frames": 27607757,
"total_tasks": 49596,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 500,
"fps": 15,
"splits": {
"train": "0:95584"
},
"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/cadene/droid_1.0.1_v30.droid_success_high_quality
DROID Success (High-Quality Extrinsics)
Subset of ~17k successful episodes in DROID-COMMUNITY filtered for high-quality camera extrinsics.
Ported from raw 1.0.1 data at full resolution to LeRobotDataset v3.0 format (0.33 TiB | 3.6k inodes) with extra annotation from KarlP/droid.
Your browser does not support the video tag.
Dataset Structure
The external cameras are assigned to left and right views depending on the episode. For their extrinsics… See the full description on the dataset page: https://huggingface.co/datasets/jnogga/droid_success_high_quality.droid_1.0.1This 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": 500,
"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/aractingi/droid_1.0.1.droid_split_00droid_3dThis dataset was created using LeRobot.
IMPORTANT: Standard LeRobot (huggingface/lerobot) does not currently support depth videos or point clouds. To load this dataset, you must use the modified codebase:
ZibinDong/lerobotdataset3d, which adds full decoding support for depth videos (H.265/H.264/FFV1) and quantized point clouds.
Dataset Description
droid_3d is a large-scale robot manipulation dataset collected with the DROID data collection platform. It contains multi-view RGB… See the full description on the dataset page: https://huggingface.co/datasets/ZibinDong/droid_3d.droid_success
DROID Success Episodes
Successful episodes in DROID-COMMUNITY. Episodes with inconsistent data or lacking language instructions were discarded, leaving ~89% of all successful episodes.
Ported from raw 1.0.1 data at full resolution to LeRobotDataset v3.0 format (1.04 TiB | 11.3k inodes) with extra annotation from KarlP/droid.
Your browser does not support the video tag.
For a subset of episodes with more reliable camera extrinsics, see… See the full description on the dataset page: https://huggingface.co/datasets/jnogga/droid_success.droid_720p
DROID 1.0.1 — native 720p, packed (LeRobot v3.0)
A native 1280×720 re-encode of the DROID robot-manipulation
dataset in LeRobot v3.0 format, with videos packed into ~500 MB files like the original
lerobot/droid_1.0.1. It reuses droid_1.0.1's
exact tabular data (states / actions / camera extrinsics / task instructions) and replaces the
low-resolution (320×180) video with full 1280×720 frames decoded from the original DROID raw MP4s
(left-eye mono per camera).
What's… See the full description on the dataset page: https://huggingface.co/datasets/XinxuanLuM/droid_720p.droid_1.0.1custom_droid_datasetThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "panda",
"total_episodes": 42,
"total_frames": 11502,
"total_tasks": 12,
"total_videos": 84,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 15,
"splits": {
"train": "0:42"
},
"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/mousecpn/custom_droid_dataset.droid_1.0.1This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "Franka",
"total_episodes": 95617,
"total_frames": 27618651,
"total_tasks": 49611,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 500,
"fps": 15,
"splits": {
"train": "0:95617"
},
"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/ygtxr1997/droid_1.0.1.droid_1.0.1
DROID 1.0.1 for pi05 training: successful, non-idle segments
lerobot/droid_1.0.1 rewritten twice, videos untouched. First
augment_droid_delta_ee_gripper_events.py (lerobot_policy_framepick) added action.delta_ee,
observation.extrinsics.static1/static2/wrist1 and the observation.gripper.time_* columns;
the flat action is the delta-EE command and the flat observation.state the EE state. Then
benchmarks/robolab/augment_droid_dataset.py (this revision) mirrored the data pipeline of… See the full description on the dataset page: https://huggingface.co/datasets/dgrachev/droid_1.0.1.droid-failure-sampled
DROID Robot Manipulation Dataset (Sampled)
数据集概述
这是从 DROID 1.0.1 数据集中采样的机器人操作失败案例子集。
总样本数: 2064
数据类型: failure
采样策略: balanced
视频格式: MP4, 60fps, 1280x720
数据集结构
hg_data/
├── videos/ # 视频文件
│ ├── 0000.mp4
│ ├── 0001.mp4
│ └── ...
├── metadata/ # 元数据文件
│ ├── 0000.json
│ ├── 0001.json
│ └── ...
├── dataset_info.json # 数据集总体信息
└── README.md # 本文件
任务类别分布
任务类别
数量
占比
Open a drawer and take some items out
4… See the full description on the dataset page: https://huggingface.co/datasets/JiaaqiLiu/droid-failure-sampled.SemanticVLA-TraceX-240K-DROID
SemanticVLA TraceX 240K · DROID
🎉 Accepted to CVPR 2026.
✍️ Fei Ni¹, Zhuo Chen², Yifu Yuan³, Zibin Dong³, Xianze Yao³, Shan Luo², Jianye Hao³, Jiankang Deng¹†, Stefanos Zafeiriou¹†
🏫 ¹Imperial College London ²King's College London ³Tianjin University
✉️ Primary contact: f.ni@imperial.ac.uk
The DROID component of TraceX-240K — the trace-annotated trajectory corpus introduced in SemanticVLA. This package is a LeRobot v3.0 repack of DROID · Franka · Open-X-Embodiment DROID… See the full description on the dataset page: https://huggingface.co/datasets/spikefly/SemanticVLA-TraceX-240K-DROID.open_drawer_droidThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "panda",
"total_episodes": 10,
"total_frames": 2646,
"total_tasks": 6,
"total_videos": 20,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 15,
"splits": {
"train": "0:10"
},
"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/mousecpn/open_drawer_droid.droid_1.0.1_v30_merged
Leyo/droid_1.0.1_v30_merged
In-place merged per chunk.
One row per episode in each data/chunk-XXX/file-YYY.parquet.
Timestamp-level columns are lists; episode-level fields are meta__*.
Join key: episode_index.
Generated with Polars (lazy/streaming), parallelized across files.
Parquet compression: zstd.
droid-newLKLAB_DROIDdroid_1.0.1_v30_compact_5This 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": 500,
"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/cadene/droid_1.0.1_v30_compact_5.droid_so101_demoThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"observation.state": {
"dtype": "float32",
"shape": [
6
],
"names": [
"shoulder_pan.pos",
"shoulder_lift.pos",
"elbow_flex.pos",
"wrist_flex.pos",
"wrist_roll.pos"… See the full description on the dataset page: https://huggingface.co/datasets/Long27/droid_so101_demo.