We provide a detailed overview of the Intel® Distribution of OpenVINO™ toolkit. We describe the toolkit components (Inference Engine, Model Optimizer, OpenCV, etc.), their capabilities and applications. The practical use of the OpenVINO™ toolkit is represented on the example of semantic segmentation problem. We consider step-by-step implementation of the problem solution based on deep learning approach. The Inference Engine component is used for high-performance inference of deep models. We demonstrate synchronous and asynchronous inference modes, and implement the first one.
Hardware and software requirements
- 3d generation of Intel® Core™ processors or higher
- Ubuntu 16.04 or higher (LTS) 64-bit
- Centos 7.4 or higher 64-bit
- Microsoft Windows 10 64-bit
- Python 3.*
- Intel® Distribution of OpenVINO™ toolkit R5 or higher [https://software.intel.com/en-us/openvino-toolkit]
* If possible, tutorial organizers will provide an access to servers with preinstalled software. In this case, only Internet access is required.