Are there instructions for training so we can contribute our own models? I've been wanting to play with ML and this seems like a fun way to do it. I read the write up and briefly looked at the repos and didn't see training instructions.