CLARU
sphita2026apache-2.0

NVIDIA Official Blogs Tech Dataset

A structured dataset comprising 2,269 articles published on the official NVIDIA Blog from 2014-2026, distributed in ZSTD-compressed Parquet format with multimodal assets including featured images, inline diagrams, and embedded videos.

Downloads50
Episodes2269

Why This Matters for Physical AI

While primarily a text and blog dataset, this collection contains 229 robotics-related articles from NVIDIA covering robotics research, applications, and infrastructure developments relevant to understanding the broader AI and hardware landscape supporting robotics research.

Technical Profile

Modalities
languageimagevideo
Episodes
2269
Data Format
parquet
Annotation Types
language_instructionsseo_metadatacategory_labelstopic_tags
License
apache-2.0
Part of the NVIDIA Official Blogs Tech Dataset family

Access

View on HuggingFace

Need custom language data?

Claru builds purpose-built datasets for any environment applications with dense human annotations and quality assurance.

Request a Sample Pack

Related Datasets

FineART: Fine-Grained Annotated Robotic Trajectory Dataset

A bimanual manipulation dataset with 40,543 teleoperated episodes (1,718 hours) featuring dense, human-written subtask annotations across 151 tasks collected on a Trossen Stationary AI robot. The dataset includes 533,913 subtask labels enabling training of hierarchical vision-language-action policies.

rgbproprioceptionlanguage
806K downloadsOct 2026other

PrimeBot Household Bimanual Manipulation Challenge Dataset

Real-world teleoperation dataset for dual-arm household manipulation tasks with frame-accurate language annotations, including over 12 manipulation tasks in household scenarios.

rgbproprioceptionlanguage
236K downloadsSep 2026cc-by-sa-4.0

OmniAction

A large-scale multimodal dataset for proactive robot manipulation comprising 141,162 episodes across 112 skills and 748 objects, enriched with audio, visual, and contextual instruction data for cross-modal intention recognition.

rgbaudiolanguage
210K downloadsMar 2026cc-by-nc-4.0

SharpaDex v1.0

A multimodal dataset of 28,993 real-world teleoperated demonstrations for bimanual dexterous manipulation across 59 tasks, comprising 32,396,172 frames (approximately 300 hours) collected from a bimanual robotic system with two 7-DoF arms, dexterous hands, and tactile sensors.

rgbtactileproprioceptionforce_torque+1
169K downloadsOct 2026cc-by-4.0

Egocentric-100K

The largest dataset of manual labor with 100,405 hours of egocentric video from head-mounted fisheye cameras, featuring state-of-the-art hand visibility and active manipulation density.

rgbvideo
148K downloadsFeb 2026apache-2.0

HiFi-UMI-2K: High-Fidelity Robot-Free Manipulation Data

A large-scale dataset of 2,000 hours of high-fidelity bimanual human manipulation demonstrations collected robot-free using the HiFi-UMI capture system, designed for action-grounded supervision for manipulation policy pre-training and deployment without requiring a robot during data collection.

rgbimuproprioceptionlanguage
140K downloadsJul 2026cc-by-4.0