Are there any how-to guides to getting something like this running? Or are there missing pieces of closed-source software needed? I'm interested in the theory but it helps to be able to try it out.
[0] http://kastnerkyle.github.io/posts/bad-speech-synthesis-made...