video1.0<iframe src="https://www.loom.com/embed/314be2433b0a4232a44c35d01332499f" frameborder="0" width="1280" height="960" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>9601280Loomhttps://www.loom.com9601280https://cdn.loom.com/sessions/thumbnails/314be2433b0a4232a44c35d01332499f-00001.gif1253.577Hi, my name is Tu and in this video, I will be presenting GIST-SUM, an approach to summarizing GitHub repositories. We all know that GitHub is the most popular platform for storing code, but for large projects, the readme file can be lengthy and time-consuming to read. That's why we need to summarize the repository. Our approach is based on text summarization techniques, specifically attractive summarization, which involves selecting important key phrases from the readme file to produce a summary without generating any new text. We used Transformers as our classification engine and fine-tuned it on our own dataset collected from GitHub. We compared our approach with ITAP and achieved better performance in all metrics. Our solution helps reduce time and effort for developers when exploring a new project for the first time. We plan to evaluate GIST-SUM with more data collected from GitHub in the future.