Paper Code Blog Blog InstructCoder (CodeInstruct): Empowering Language Models to Edit Code Updates May 23, 2023: Paper, code and data released. Overview InstructCoder is the first dataset designed to adapt LLMs for general code editing. It consists of 114,239 instruction input output triplets and covers multiple distinct code editing scenarios, generated by ChatGPT. LLaMA 33B finetuned on InstructCoder performs on par with ChatGPT on a real world test set derived from GitHub commits. In the ever evolving world of software development, efficient and effective code editing plays a pivotal role. As the demands for faster development cycles and increased productivity continue to rise, harnessing the power of Large Language Models in code editing has emerged as a game changer. This project aims to revolutionize code editing, empowering developers with intelligent tools and techniques. Though the community has studied extensively on coding with LLMs, our experiments demonstrate that current LLMs still struggle for code editing. To this end, we curate a code editing dataset, dubbed InstructCoder, for improving and evaluating code editing abilities of LLMs. InstructCoder is an instructiona…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy