datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
VideoHallu
VideoHallu: Evaluating and Mitigating Multi-modal Hallucinations for Synthetic Videos
Zongxia Li*, Xiyang Wu*, Guangyao Shi, Yubin Qin, Hongyang Du, Tianyi Zhou, Dinesh Manocha, Jordan Lee Boyd-Graber
[📖 Paper] [🤗 Dataset] [🌍Website]
👀 About VideoHallu
Synthetic video generation has gained significant attention for its realism and broad applications, but remains prone to violations of common sense and physical laws. This highlights the need for reliable abnormality… See the full description on the dataset page: https://huggingface.co/datasets/IntelligenceLab/VideoHallu.SenseXperience_WristCam_SampleData
SenseXperience Raw MCAP Sample Data
Raw capture episodes from SenseXperience (IO-AI): 12 human motion episodes in ROS 2 MCAP, with 4× compressed video + head IMU. Format details: data format reference.
Item
Value
Episodes
12
Date
2026-07-13
Format
ROS 2 MCAP
Duration
~63–100 s / episode
Modalities
4 cameras (MJPEG) + head IMU (~120 Hz)
Layout
episode_<id>_yyyy_mm_dd_hh_mm_ss/
├── *_mcap_0.mcap # ROS 2 MCAP bag
├── metadata.yaml… See the full description on the dataset page: https://huggingface.co/datasets/io-intelligence/SenseXperience_WristCam_SampleData.realman_aidal_desktop_cleanupThe dataset was collected and open-sourced by IO Intelligence, and exported in the LeRobot format provided by the IO Data Platform.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "custom_arm",
"total_episodes": 1099,
"total_frames": 246816,
"total_tasks": 323,
"total_videos": 4396,
"total_chunks": 2,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:1099"},
"data_path":… See the full description on the dataset page: https://huggingface.co/datasets/io-intelligence/realman_aidal_desktop_cleanup.FoldingTShirt_DualArxR5a_Samples
FoldingTShirt_DualArxR5a_Samples
100 real-robot teleoperation episodes for “Fold the T-shirt on the table.” on a DualArxR5a dual-arm robot. Format: raw MCAP (ROS 2 / rosbag2).
Source
Collected with TeleXperience, IO-AI’s product for real-robot teleoperation and data collection. An operator drives the robot; TeleXperience writes time-aligned RGB, joint commands, joint states, gripper targets, and end-effector poses to MCAP.
Product page:… See the full description on the dataset page: https://huggingface.co/datasets/io-intelligence/FoldingTShirt_DualArxR5a_Samples.Emotion.Intelligenceegoproactive-synth-annotations
EgoProactive synthetic proactive annotations
Everything produced by the annotation and synthesis pipelines for the AI Wearables Challenge 2026 EgoProactive
Dense timestamped proactive walkthroughs generated with the ambient agent
(orchestrator deepseek/deepseek-v4-flash-0731 + vision Qwen3.6-27B),
using the held-out-validated dense policy (setup-phase coverage, repetition-collapse,
fire-at-onset) and a -0.5s onset correction at chunk-binning.
set
clips
median events/clip… See the full description on the dataset page: https://huggingface.co/datasets/ambient-intelligence-labs/egoproactive-synth-annotations.so101_stack_cupsThis dataset was created using LeRobot.
Dataset Description
SO-101 (so_follower) teleoperation dataset for stacking cups. Contains 660 episodes / 188035 frames at 30 FPS, with three RGB cameras (observation.images.front, observation.images.top, observation.images.wrist) and 6-DoF joint state/action. Task prompt: "Stack the cups". Stored in LeRobot v3.0 format.
Homepage: https://huggingface.co/datasets/io-intelligence/so101_stack_cups
Paper: [More Information Needed]… See the full description on the dataset page: https://huggingface.co/datasets/io-intelligence/so101_stack_cups.WipeTable_DualArxR5a_TeleXperienceThis dataset was created using LeRobot.
Dataset Description
73 real-robot teleoperation episodes for “Wipe the table.” on a DualArxR5a dual-arm robot. Format: LeRobot v3.0 (30 Hz parquet + H.264 videos).
Collected with TeleXperience, IO-AI’s product for real-robot teleoperation and data collection.
Task / language prompt: Wipe the table.
Robot: DualArxR5a (bimanual, parallel-jaw grippers)
Frames: 629523 at 30 Hz
Cameras: camera_high (overhead), camera_low (lower scene)… See the full description on the dataset page: https://huggingface.co/datasets/io-intelligence/WipeTable_DualArxR5a_TeleXperience.EgoSafetyBench
EgoSafetyBench — Dataset
Ego-view (chest-camera) physical-safety benchmark for evaluating vision-language
models as runtime safety guards for humanoid robots. Each clip is a short,
physically grounded moment; a guard must classify whether the unfolding action is
safe or unsafe, and whether an in-scene channel (a sign, label, or screen) is
misleading.
Two-axis taxonomy
Every video is labeled along two independent axes:
Situational family — what the physical scene… See the full description on the dataset page: https://huggingface.co/datasets/AIM-Intelligence/EgoSafetyBench.cyclo_intelligence_eefpose_test_lerobot_v21
Task_11_11_MCAP
Created with Cyclo Intelligence by ROBOTIS.
Task_cyclo_intelligence_pickplacetip_lerobot_v21
Task_0001_pickplacetip_MCAP
Created with Cyclo Intelligence by ROBOTIS.
cyclo_intelligence_test_dataset_lerobot_v2.1
Task_10_10_MCAP
Created with Cyclo Intelligence by ROBOTIS.
ChenLong_Embodied_Intelligence_Dataset
ChenLong Embodied Intelligence Dataset
本仓库用于统一管理辰龙机器人实习中的数据集、模型权重、训练结果和说明文档。后续新增不同任务、采集批次、模型版本或实验资源时，都放在这里统一维护。
当前目录
embodied_dataset/：具身智能采集数据集，采用 LeRobot v3.0 结构，包含 data/、meta/、videos/。
yolo_dataset/：YOLO 目标检测数据、模型权重、训练参数和评估结果，当前包含 blue_bucket_yolov8/。
待新增新的数据集或模型。
新增数据集要求
具身数据优先采用 LeRobot v3.0 格式：meta/info.json、meta/stats.json、tasks、episodes、逐帧 Parquet 数据和按相机划分的视频。新增数据集至少写清：
任务：任务文本、目标物、成功标准、失败标准。
硬件：机器人型号、自由度、夹爪、相机位置、分辨率、FPS。… See the full description on the dataset page: https://huggingface.co/datasets/vvzc/ChenLong_Embodied_Intelligence_Dataset.cyclo_intelligence_subtask_failed_dataset_test_lerobot_v21
Task_1_subtask_failed_dataset_test_MCAP
Created with Cyclo Intelligence by ROBOTIS.
arx_pick_block
ARX Pick Block Dataset
This dataset was created using LeRobot and contains dual-arm robotic manipulation demonstrations for pick-and-place tasks.
Task Overview
This dataset contains demonstrations of a dual-arm robotic manipulation task where the robot picks up a block from a desk and places it into a plate.
Task Details
Task Objective: Pick up the block from the desk and place into plate
Operational Objects: Block
Operation Duration: Each operation takes… See the full description on the dataset page: https://huggingface.co/datasets/io-intelligence/arx_pick_block.benchgen_lerobot_dataset_filteredeval_act_so100_testThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "so100",
"total_episodes": 10,
"total_frames": 11903,
"total_tasks": 1,
"total_videos": 10,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"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/Creative-Intelligence/eval_act_so100_test.cyclo_intelligence_test_dataset_lerobot_v3.0
Task_10_10_MCAP
Created with Cyclo Intelligence by ROBOTIS.
cup2This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "aloha",
"total_episodes": 3,
"total_frames": 2679,
"total_tasks": 1,
"total_videos": 6,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:3"
},
"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/Creative-Intelligence/cup2.aloha_testThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "aloha",
"total_episodes": 2,
"total_frames": 1320,
"total_tasks": 1,
"total_videos": 6,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:2"
},
"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/Creative-Intelligence/aloha_test.so100_testThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "so100",
"total_episodes": 2,
"total_frames": 1192,
"total_tasks": 1,
"total_videos": 2,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:2"
},
"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/Creative-Intelligence/so100_test.cyclo_intelligence_lerobot_test
merge_test_1
Created with Cyclo Intelligence by ROBOTIS.
cyclo_intelligence_lerobot_dataset_test
merge_test_1
Created with Cyclo Intelligence by ROBOTIS.
cyclo_intelligence_omx_test_v21
Task_1_omx_test_MCAP
Created with Cyclo Intelligence by ROBOTIS.
Visual-Intelligence
Visual-Intelligence
🔗 Links
💾 Github Repo
🤗 HF Dataset
📑 Blog
📖 Dataset Introduction
Dataset Schema
id: Unique sample identifier.
input: Ordered list describing the input context.
type: Either "image" or "text".
content: For "image", a relative path to the first-frame image. For "text", the prompt text.
output: Generated candidates and final selections by model.
veo3: Relative paths to videos generated by the VEO3 pipeline.
framepack:… See the full description on the dataset page: https://huggingface.co/datasets/Entroplay/Visual-Intelligence.cyclo_intelligence_test_lerobot_1
merge_test_1
Created with Cyclo Intelligence by ROBOTIS.
Video-Intelligence-resultssession_2026-07-15_11-34-45Video-Intelligence-results-new