victor-betus/github-issues-dataset

该数据集包含从GitHub上排名前100的仓库中收集的114,073个GitHub问题，设计用于问题分类、严重性/优先级预测以及AI/ML训练。具体来说，它可用于AI/ML训练（如微调模型进行问题分类和优先级排序）、自然语言处理（NLP）分析软件开发讨论，以及错误严重性预测（训练模型将问题分类为关键、主要或次要）。数据集以Parquet格式存储，包含id、repo、title、body、labels、priority和severity等列。

This dataset contains 114,073 GitHub issues collected from the top 100 repositories on GitHub. It is designed for issue classification, severity/priority prediction, and AI/ML training. Specifically, it is useful for AI/ML training (e.g., fine-tuning models for issue classification and prioritization), Natural Language Processing (NLP) analysis of software development discussions, and bug severity prediction (training models to classify issues as Critical, Major, or Minor). The dataset is stored in Parquet format and includes columns such as id, repo, title, body, labels, priority, and severity.