Yes, it was. From their site: "It has been trained on a selection of English language and source code from publicly available sources, including code in public repositories on GitHub."