Datasets
Robot demonstrations, egocentric video and speech corpora for training physical AI, each linked to its home on Hugging Face.
Filters
Task
Data
HardwareCompute boards show an estimated fit from parameter counts, not a benchmark. Devices and robots show documented associations.
Embodiment
Results
15 datasets
- LElerobot/svla_so101_pickplaceRobot demonstrationsApache 2.0May 2025
Fifty real pick-and-place episodes from a low-cost SO-100/SO-101-class arm, recorded at 30 fps with a side camera and a top-down camera, and used as the SmolVLA fine-tuning example.
- HUHuggingFaceVLA/community_dataset_v1Robot demonstrationsApache 2.0Aug 2025
A filtered collection of 128 SO-100 datasets from 55 community contributors (about 11k episodes and 5.1M frames), with hand-cleaned task descriptions, released as the pretraining subset for SmolVLA.
- LElerobot/pushtRobot demonstrationsMITMar 2024
About 200 human demonstrations in a 2D simulation where a round pusher slides a T-shaped block onto a target outline, stored in LeRobot v3 format.
- LElerobot/aloha_sim_transfer_cube_humanRobot demonstrationsMITMar 2024
Fifty human-teleoperated simulated episodes of a two-arm ALOHA robot handing a cube from one gripper to the other, recorded at 50 fps.
- LElerobot/droid_1.0.1Robot demonstrationsApache 2.0Jul 2025
LeRobot v3 conversion of DROID: about 95k real Franka teleoperation episodes recorded in many different scenes, with around 50k distinct language instructions.
- LElerobot/liberoRobot demonstrationsApache 2.0Jan 2026
LeRobot-format copy of the LIBERO simulated Franka Panda benchmark, with about 1.7k demonstrations across its 40 language-conditioned tasks.
- ICIPEC-COMMUNITY/bridge_orig_lerobotRobot demonstrationsApache 2.0Feb 2025
LeRobot v2.0 conversion of BridgeData V2: about 53k WidowX tabletop and toy-kitchen episodes paired with roughly 20k distinct language instructions.
- JXjxu124/OpenX-EmbodimentRobot demonstrationsCC BY 4.0Oct 2023
Unofficial community mirror of the Open X-Embodiment collection: about 55 robot sub-datasets repackaged as tar shards, together covering more than a million real trajectories.
- NVnvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-SimRobot demonstrationsCC BY 4.0Mar 2025
NVIDIA's simulated trajectory sets for post-training GR00T, covering bimanual Panda, Fourier GR1 humanoid tabletop, single-arm Panda kitchen, BEHAVIOR R1 Pro household and Unitree G1 tasks, each stored as a LeRobot-style folder.
- AWagibot-world/AgiBotWorld-AlphaRobot demonstrationsCC BY-NC-SA 4.0GatedDec 2024
Real dual-arm demonstrations recorded on a fleet of AgiBot robots across 36 household, retail and industrial tasks, shipped as tar archives with a script to convert them to LeRobot format.
- UPUSC-PSI-Lab/humanoid-everydayRobot demonstrationsApache 2.0Oct 2025
Teleoperated humanoid demonstrations across more than 260 tasks in seven categories, with RGB, depth, LiDAR, tactile and IMU streams at 30 Hz, stored in LeRobot v2 format.
- BUbuilddotai/Egocentric-10KEgocentric videoApache 2.0GatedNov 2025
About 10,000 hours of 1080p head-mounted video of factory workers doing hands-on jobs, delivered as roughly 193k H.265 clips (about 16 TB) with per-clip metadata.
- OPopenslr/librispeech_asrSpeech corpusCC BY 4.0Mar 2022
Roughly 1,000 hours of read English audiobook speech with transcripts, the long-standing benchmark for English speech recognition.
- GOgoogle/fleursSpeech corpusCC BY 4.0Apr 2022
N-way parallel read-speech set in about 100 languages, used to evaluate multilingual speech recognition, language identification and speech translation.
- GOgoogle/speech_commandsSpeech corpusCC BY 4.0Mar 2022
One-second clips of short spoken words such as yes, no, stop and the digits, collected for training tiny keyword-spotting models.