Installing Intel® Distribution of OpenVINO™ Toolkit¶
Intel® Distribution of OpenVINO™ Toolkit is a comprehensive toolkit for developing applications and solutions based on deep learning tasks, such as computer vision, automatic speech recognition, natural language processing, recommendation systems, and more. It provides high-performance and rich deployment options, from edge to cloud. Some of its advantages are:
Enables CNN-based and transformer-based deep learning inference on the edge or cloud.
Supports various execution modes across Intel® technologies: Intel® CPU, Intel® Integrated Graphics, Intel® Discrete Graphics, Intel® Neural Compute Stick 2, and Intel® Vision Accelerator Design with Intel® Movidius™ VPUs.
Speeds time-to-market via an easy-to-use library of computer vision functions and pre-optimized kernels.
Compatible with models from a wide variety of frameworks, including TensorFlow, PyTorch, PaddlePaddle, ONNX, and more.
OpenVINO 2022.3, temporarily, does not support the VPU devices. The feature will be re-implemented with the next update. Until then, continue using a previous release of OpenVINO, if you work with VPUs.
Check out the OpenVINO Download Page
OpenVINO installation package is distributed as two options: OpenVINO Runtime and OpenVINO Development Tools.
OpenVINO Runtime contains the core set of libraries for running machine learning model inference on processor devices.
OpenVINO Development Tools is a set of utilities for working with OpenVINO and OpenVINO models. It includes the following tools:
Post-Training Optimization Tool
Accuracy Checker and Annotation Converter
Model Downloader and other Open Model Zoo tools
Option 1. Install OpenVINO Runtime and OpenVINO Development Tools (recommended)¶
The best way to get started with OpenVINO is to install OpenVINO Development Tools, which will also install the OpenVINO Runtime Python package as a dependency. Follow the instructions on the Install OpenVINO Development Tools page to install it.
For developers working in Python, OpenVINO Development Tools can easily be installed using PyPI. See the For Python Developers section of the Install OpenVINO Development Tools page for instructions.
For developers working in C++, the core OpenVINO Runtime libraries must be installed separately. Then, OpenVINO Development Tools can be installed using requirements files or PyPI. See the For C++ Developers section of the Install OpenVINO Development Tools page for instructions.
Option 2. Install OpenVINO Runtime only¶
OpenVINO Runtime may also be installed on its own without OpenVINO Development Tools. This is recommended for users who already have an optimized model and want to deploy it in an application that uses OpenVINO for inference on their device. To install OpenVINO Runtime only, follow the instructions on the Install OpenVINO Runtime page.
The following methods are available to install OpenVINO Runtime:
Linux: You can install OpenVINO Runtime using APT, YUM, archive files or Docker. See Install OpenVINO on Linux.
Windows: You can install OpenVINO Runtime using archive files or Docker. See Install OpenVINO on Windows.
macOS: You can install OpenVINO Runtime using archive files or Docker. See Install OpenVINO on macOS.
Option 3. Build OpenVINO from source¶
Source files are also available in the OpenVINO Toolkit GitHub repository. If you want to build OpenVINO from source for your platform, follow the OpenVINO Build Instructions.
Still unsure if you want to install OpenVINO toolkit? Check out the OpenVINO tutorials to run example applications directly in your web browser without installing it locally. Here are some exciting demos you can explore:
Follow these links to install OpenVINO: