Build and test image generation on Windows client

[Savokr]

After investigation, I came up to the idea of using VoltaML as inference engine and issuing local network requests for getting wanted images instead of running shell commands for each new request. In the client then, there should be just a local IP address for running server with VoltaML app.