Code for the paper "Efficient Training of Language Models to Fill in the Middle" - View it on GitHub