OpenVINO™ is an open source toolkit for optimizing and deploying AI inference

A repository for storing models that have been inter-converted between various frameworks. Supported frameworks are TensorFlow, PyTorch, ONNX, OpenVINO, TFJS, TFTRT, TensorFlowLite (Float32/16/INT8), EdgeTPU, CoreML.

📚 Jupyter notebook tutorials for OpenVINO™

Train, Evaluate, Optimize, Deploy Computer Vision Models via OpenVINO™

Open-source Linux performance suite for engineers—profiling and tuning workloads and system configurations.

🤖 Provide 70+ ready-to-use, platform-agnostic AI agent skills for improving tasks like code review, security, and data analysis.

Libva is an implementation for VA-API (Video Acceleration API)

Concise Binary Object Representation (CBOR) Library

61 ready-to-use AI agent skills for go-to-market teams, installable in Claude Code with a bootstrap script and onboarding flow.

Cyber threat intelligence and OSINT analysis toolkit. Runs structured investigations and delivers analyst-grade intelligence products with sourced, trust-scored findings. Use for OSINT and CTI cases, digital-footprint and exposure review, domain/subdomain/DNS/certificate recon, web-infrastructure pivoting (favicon hashes, tracker IDs, TLS certs, phishing-kit fingerprinting, campaign clustering), username/email/phone enumeration, breach and infostealer-log triage, image forensics, geolocation, crypto-wallet and IBAN/bank-account tracing, darknet search, M365/Azure and SaaS tenant recon, China/Sinophone recon (ICP filings, PRC corporate registries, Baidu/FOFA/Quake/ZoomEye), vulnerability and ransomware lookup, threat modeling, PII redaction, and structured reporting. Commands include /case, /sweep, /query, /webpivot, /username, /phone, /email-deep, /breach-deep, /icp, /cn-corp, /iban, /stealer-log, /exposure, /threat-model, /report, /brief, /redact, /apikeys.

Real-time 3D multi-person pose estimation demo in PyTorch. OpenVINO backend can be used for fast inference on CPU.

This repository intends to enable autonomous drone delivery with the Intel Aero RTF drone and PX4 autopilot. The code can be executed both on the real drone or simulated on a PC using Gazebo. Its core is a robot operating system (ROS) node, which communicates with the PX4 autopilot through mavros. It uses SVO 2.0 for visual odometry, WhyCon for visual marker localization and Ewok for trajectoy planning with collision avoidance.