I have got a question concerning the development and inference of Neural Networks with the help of Open Vino Toolkit.
The goal of the student research team, whom I am a part of, is to run a neural network on FPGA in order to allow emotion detection. We are due to use Open Vino GX Starter Kit (Cyclone V device). The problem is the following: the other crucial blocks of the project work with API and software available for Windows 10, but, as from the installation guide, the FPGA support is present in Ubuntu, Cent Os and Poky Jethro OS only.
Is it possible to run the Model Optimizer and Inference Engine on Ubuntu, for instance, and communicate with the board via PCIe from the user application on Windows (without any reference to the Open Vino Toolkit)? Do we get the overall concept right? If not, is there any suggested workaround?
Thank you in advance for the response.
Hi Liliia Kudelina,
Thanks for reaching out. At this time, the OpenVINO toolkit for windows does not support the FPGA devices. Please also note that the OpenVINO toolkit for Linux (FPGA) only supports the following products as mentioned in the system requirements (FPGA Section).
- Intel® Arria® 10 FPGA GX development kit
- Intel® Programmable Acceleration Card with Intel® Arria® 10 GX FPGA operating systems