Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

awesome-openvino

A curated list of OpenVINO based AI projects
https://github.com/openvinotoolkit/awesome-openvino

Last synced: 17 days ago
JSON representation

Further resources:

Table of content

AI Audio

  • Whisper OpenVINO
  • Sangeet Guru - A music generation app where users input a music style description to get custom audio tracks.
  • OpenVINO™ AI Plugins for Audacity® - A set of AI-enabled effects, generators, and analyzers for Audacity® such as Music Stem Separation, Noise Suppression, Transcription, and Music Generation.

AI Computer Vision

  • LidarObjectDetection-PointPillars
  • Pedestrian fall detection - Pedestrian fall detection. Deploying PP-Human based on OpenVINO C # API
  • Clip-Chinese - Chinese image-text similarity matching tasks, leverage OpenVINO and the Towhee embedding library.
  • Visioncom
  • BMW-IntelOpenVINO-Detection-Inference-API - This is a repository for an object detection inference API using OpenVINO, supporting both Windows and Linux operating systems
  • yolov5_export_cpu - The project provides documentation on exporting YOLOv5 models for fast CPU inference using Intel's OpenVINO framework
  • Image Processing with OpenVINO
  • Implementing GAN with OpenVINO
  • RapidOCR
  • Cigarette Detection - The project begins by YOLOv8-pose detecting human body positions and extracting skeletal information from images. Based on the skeletal poses, it assesses the elbow angles and the distance between hands and mouths for each individual. If successful, the RTDETR model is employed to detect cigarettes at the mouth zone.
  • FastSAM_Awesome_OpenVINO - The Fast Segment Anything Model(FastSAM) is a CNN Segment Anything Model trained by only 2% of the SA-1B dataset published by SAM authors. The FastSAM achieve a comparable performance with the SAM method at 50× higher run-time speed.
  • Computer Vision Models As Service - implements different Computer Vision Deep Learning Models as a service.
  • Dance-with: Dance with your friends with the right pose! - Dance-with corrects your dance posture using multi-person OpenPose, 2D pose estimation Deep Learning model.
  • Target-Person-Tracking-System - Integration of face recognition and person tracking.
  • Metin2 Bot - bots for video game Metin2.
  • Machine control - industrial machine surveillance system designed to help increase efficiency of processes.
  • MeetingCam - Run your AI and CV algorithms in online meetings such as Zoom, Meets or Teams!
  • Virtual-Tryon - Use AI to try on clothes with your pictures.
  • DepthAI Experiments - A collections of projects done with DepthAI.
  • Group Pose - A Simple Baseline for End-to-End Multi-person Pose Estimation.
  • Frigate - NVR With Realtime Object Detection for IP Cameras.
  • CGD OpenVINO Demo - Efficient Inference and Quantization of CGD for Image Retrieval.
  • Risk package detection - Threat Detection and Unattended Baggage Detection with Associated Person Tracking.
  • YOLOv7-Intel - Object Detection For Autonomous Vehicles.
  • Cerberus - Dog Breed Classification and Body Localization.
  • Criminal Activity recognition - Criminal Activity Video Surveillance.
  • RapidOCR on OpenVINO GPU - A modified verison of RapidOCR to support OpenVINO GPU.
  • Yolov9 with OpenVINO - C++ and python implementation of YOLOv9 using OpenVINO
  • OpenVINO-Deploy - A repository showcasing the deployment of popular object detection AI algorithms using the OpenVINO C++ API for efficient inference.
  • VisionGuard - A desktop application designed for AI PCs to combat eye strain through real-time gaze tracking, developed during GSoC 2024 under the OpenVINO Toolkit. Built on OpenVINO's gaze estimation demo, VisionGuard offers customizable break reminders, screen time analytics, and multi-device support (CPU/GPU/NPU). It features an intuitive UI with system tray integration, leveraging OpenVINO, Qt, and OpenCV for efficient, privacy-focused local processing.
  • Project Babble - Mouth tracking project designed to work with any existing VR headset.
  • OpenVINO Tennis Posture - Deciphering Tennis Posture with Artificial Intelligence

Educational

Frameworks

  • Keras 3 - Keras 3 is a multi-backend deep learning framework, with support for JAX, TensorFlow, PyTorch, NumPy and OpenVINO. User can switch on OpenVINO backend for models inference using Keras API.
  • OpenVINO torch.compile backend - Run PyTorch models natively on OpenVINO using torch.compile, without exporting to ONNX or IR. Compatible with LLMs, Diffusers, Stable Diffusion, and Intel NPUs via the TorchDynamo backend.
  • ExecuTorch OpenVINO Backend - OpenVINO backend for ExecuTorch enabling optimized PyTorch edge inference on CPU, GPU, and NPU targets.
  • LangChain OpenVINO integration - Officially supported LangChain backend for running LLMs, embeddings, and rerankers with OpenVINO, suitable for RAG and Agent pipelines on CPU, GPU, and NPU.
  • langchain-openvino - Community project extending LangChain with OpenVINO support for LLMs, embeddings, Whisper, and Vision-Language Models.
  • Optimum Intel - Hugging Face integration layer for OpenVINO supporting Transformers, Diffusers, and SentenceTransformers with INT4/INT8 quantization via NNCF.

Generative AI

  • OpenVINO Code - VSCode extension for AI code completion with OpenVINO - VSCode extension for helping developers writing code with AI code assistant.
  • stable_diffusion.openvino - This GitHub project provides an implementation of text-to-image generation using stable diffusion on Intel CPU or GPU. It requires Python 3.9.0 and is compatible with OpenVINO.
  • Fast SD - FastSD CPU is a faster version of Stable Diffusion on CPU. Based on Latent Consistency Models and Adversarial Diffusion Distillation.Read blog post about Fast Stable Diffusion on CPU using FastSD and OpenVINO.
  • OpenVINO™ AI Plugins for GIMP - Provides a set of OpenVINO based plugins that add AI features to GIMP (GNU IMAGE
  • Enhancing Customer Service with Real-Time Sentiment Analysis: Leveraging LLMs and OpenVINO for Instant Emotional Insights - The integration of LLMs with sentiment analysis models, further optimised by OpenVINO.
  • OV_SD_CPP - The pure C++ text-to-image pipeline, driven by the OpenVINO native API for Stable Diffusion v1.5 with LMS Discrete Scheduler.
  • QuickStyle - A simple stylizing app utilizing OpenVINO to stylize common objects in images.
  • QuickPainter - A simple inpainting app utilizing OpenVINO to remove common objects from images.
  • BlurAnything - An adaptation of the excellent Track Anything project which is in turn based on Meta's Segment Anything and XMem.
  • Stable Diffusion 2.1 on Intel ARC - A simple and easy-to-use demo to run Stable Diffusion 2.1 for Intel ARC graphics card based on OpenVINO.
  • AI Video Builder - Make videos with AI images from YouTube videos.
  • MED-LLM-BR-OpenVINO - These models were specially developed for the clinical context in Brazilian Portuguese, proving to be efficient in generating synthetic clinical data. The models are essential not only for direct applications in healthcare, but also for training larger models, overcoming the difficulty in accessing patient record data.
  • LocalAI - LocalAI is the free, Open Source OpenAI alternative. LocalAI act as a drop-in replacement REST API that’s compatible with OpenAI (Elevenlabs, Anthropic... ) API specifications for local AI inferencing.
  • AI structured data Extraction - A streamlit interface with NuExtract-1.5-tiny and openvino-genai to extract data from plain text into structured custom json formats. The Repo also gives a small tutorial on how to convert NuExtract-1.5-tiny into OpenVINO IR format.
  • StableLM-3B Chatbot - A streamlit CHATBOT interface with stablelm-zephyr-3b quantized in 4bit and optimum-intel. The Interface has a kind text streaming effect, and the number of turns are handled to not exceed the context window. The Model used is published on Hugging Face Hub and was created with the free HF Space hosting the NCCF-quantization tool.
  • Gemma2-2b AI Chat App - A beautiful Chat Interface, with interactive tuning parameters, powered by Optimum-Intel[openvino], Streamlit and the small but powerful Gemma2-2b-instruct model by Google. The model is an int4 quantized version, hosted on Hugging Face Hub.
  • LaMini Power - An experimental text based chat interface in the terminal running the LaMini-Flan-T5-248M . This is a breakthrough made possible by openvino, because encoder-decoder model could not be quantized. The LaMini model family is a highly curated herd of very small models achieving strong accuracy even with only 512 tokens of context length.
  • ComfyUI with FastSD and OpenVINO
  • Introducing Simple Image Generator - Create Amazing Images on Your Computer. No Internet, No GPU, No Subscriptions!.
  • Podman AI Lab - Podman AI Lab is an open source extension for Podman Desktop to work with LLMs (Large Language Models) on a local environment using the OpenVINO model server. Featuring a recipe catalog with common AI use cases, a curated set of open source models, and a playground for learning, prototyping and experimentation, Podman AI Lab helps you to quickly and easily get started bringing AI into your applications, without depending on infrastructure beyond your laptop ensuring data privacy and security.
  • OpenVINO GenAI - High-level GenAI pipeline library built on OpenVINO for LLM, VLM, Whisper, and Diffusion use cases. Features speculative decoding, Prompt Lookup, and agent-ready integrations with OpenVINO Model Server.
  • Stable Diffusion web UI - This is a repository for a browser interface based on Gradio library for Stable Diffusion
  • OpenVINO OpenAI API - An OpenAI-compatible API server powered by OpenVINO GenAI for efficient inference on Intel hardware. OpenAI API compatibility for easy integration with existing applications, Powered by OpenVINO for optimized inference on Intel CPUs and GPUs, Support for both streaming and non-streaming responses, Simple command-line interface for launching the server. Easy installation with pip install openvino-openai-api. The project is created with AI coding assistance. You can read the journey and steps in the following 3 articles: part 1, part 2 and part 3
  • Cascadia - Distributed LLM inference across Intel laptops, desktops, and AI PCs. Shards a model into INT4 pipeline stages across multiple machines and serves it through an OpenAI-compatible API, using OpenVINO and OpenVINO GenAI engines (including distributed speculative decoding). Ships as a single Rust binary per node with mDNS peer discovery and a doctor command that detects when OpenVINO can't see your GPU.
  • RustedVINO - Multimodal inference server written in Rust on top of OpenVINO GenAI. A single native binary serves LLM, VLM, STT, TTS, embeddings, reranking and image generation concurrently through an OpenAI-compatible API with custom extensions. Supports dynamic model loading and unloading without restart and exposes a Prometheus metrics endpoint.

Miscellaneous

Multimodal projects

Natural Language Processing

  • Japanese chatbot Youri - LLM Japanese chatbot demo program using Intel OpenVINO toolkit.
  • OpenVINO GPT-Neo - a port of GPT-Neo that uses OpenVINO.
  • Resume-Based Interview Preparation Tool - The Resume-Based Interview Preparation Tool is a software application designed to streamline the interview process by helping interviewers generate relevant and meaningful questions based on a candidate's resume or portfolio page.

openSUSE

OpenVINO API extentions

Related Communities

Robotics

  • ROS 2 OpenVINO Toolkit - Intel toolkit for integrating OpenVINO into ROS 2 pipelines, supporting object detection, segmentation, and tracking for robotics and edge AI applications.