Locate suitable training dataset

[Pgibby8]

Right now I'm leaning towards Project Gutenberg for its size, diversity, and availability http://www.gutenberg.org/

[Pgibby8]

This repository seems promising https://github.com/c-w/gutenberg