Skip to content

Datasets

Hugging Face datasets with plain-language descriptions: what's inside and what models to train on them. 4,788 in the catalog so far, with new ones added every hour.

Data type

For which models

Find a dataset for your goal

Answered by Gemini 4 · the chat is kept only in this tab

Datasetagibot-world/AgiBotWorld2026Top pick

Real-world data for AgiBot World 2026 robots

Corpus of real scenes with images and video for training and evaluating embodied AI and robotics.

  • Robotics
  • English
2.1M 115100K–1M rows
DatasetDAGroup-PKU/Open-HumanWAM-50KTop pick

50K hours of first-person hand action videos

Large multimodal corpus of human hand manipulation videos with motion annotations and descriptions for robot learning.

  • Robotics
1.5M 3610M–100M rows🔒 on request
Datasetlerobot/FineARTTop pick

Bimanual robot trajectories with subtask labels

Large teleoperation video dataset of a bimanual robot with detailed manual step labels for training robotics models.

  • Robotics
1.1M 4610K–100K rows
DatasetSharpa-Robotics/SharpaDex-v1.0Top pick

Bimanual Robot Manipulation Demonstrations

Large-scale robot teleoperation recordings across 56 tasks using head and wrist cameras.

    516.3K 63100K–1M rows
    DatasetMINT-SJTU/RW-RL-DatasetTop pick

    Real-world robot interaction video

    Large video dataset of robot interactions with the real environment for reinforcement learning.

    • English
    417K 19100K–1M rows
    DatasetHubLabs/HubEgo-100KTop pick

    First-person real manual labor video dataset

    Over 10,000 hours of real-world recordings of craftsmanship, cooking, and construction work with IMU sensor data.

    • Robotics
    • Video classification
    • Video text → text
    • English
    257.2K 931B–10B rows🔒 on request
    DatasetLingYu-Robotics/lingyu-realworld-robotics-datasetTop pick

    LingYu Robotics dual-arm teleoperation video

    139.4 hours of first-person teleoperation video recordings for training robots in object manipulation.

    • English
    242.6K 910K–100K rows
    Datasetrobbyant/robotwin-clean-and-aug-lerobot

    Robotwin video latents for WAN 2.2

    Robotwin dataset in LeRobot format with pre-extracted video latents for fine-tuning robotics models.

      217.9K 2110K–100K rows
      Datasetgenrobot2025/10Kh-RealOmin-OpenDataTop pick

      Robot Data: 13k Hours of Manipulation

      Robot manipulation video corpus: 13k hours, 5M clips, and everyday scenarios for embodied AI training.

      • Robotics
      • Reinforcement learning
      • English
      • Chinese
      196K 281over 1T rows🔒 on request
      Datasetfacebook/seamless-interactionTop pick

      Multimodal human interaction corpus

      4,000+ hours of audio and video of in-person interactions by 4,000+ participants for AI research.

        195K 223
        DatasetWorldEngineAI/WEB-DatasetTop pick

        Bimanual manipulation dataset for robotics

        A large-scale dataset of real bimanual robot video and telemetry with text descriptions for 90 everyday tasks.

          193.1K 2100K–1M rows🔒 on request
          DatasetRekaAI/RekaDaily-10k-raw

          Raw first-person everyday life video dataset

          Unedited daily and working first-person video recordings in English for training video models.

          • Video classification
          • Image → video
          • English
          187.1K 32100K–1M rows
          Datasetfacebook/show3d-datasetTop pick

          Meta 3D hand-object interaction scenes

          Multimodal dataset with video, 3D hand and object poses in real-world conditions for CVPR 2026.

          • Other
          • English
          164K 171K–10K rows
          Datasetbuilddotai/Egocentric-100KTop pick

          100k hours of manual labor egocentric video

          Largest first-person manual labor video dataset: 100,405 hours and 2 million clips.

            149.6K 1641M–10M rows🔒 on request
            Datasetconfiginc/HABITTop pick

            Human-aware robot demonstrations (HABIT)

            Large multimodal dataset of 1M to 10M examples for training robots in safe human interaction.

            • Robotics
            • English
            136.6K 91M–10M rows
            DatasetACERobotics/ACE-Data-0

            Multimodal indoor human action recordings for robotics

            Synchronized video, audio, and 3D data from real home environments for embodied AI training.

            • Robotics
            • Keypoint detection
            • Video classification
            • English
            128.8K 6010K–100K rows
            DatasetGokuScraper/seedance-2-prompts-datasets

            Seedance 2.0 prompts and videos

            Over 8,100 English and Chinese prompts together with generated videos and preview frames.

            • Text → video
            • English
            • Chinese
            124.5K 611K–10K rows