This project fine-tuned compact monolingual code language models to new programming languages using efficient techniques, with the goal of expanding capabilities and accessibility of code intelligence tools. The work shares insights from optimizing training efficiency and makes models/datasets publicly available to further AI democratization.
