nishide-dev/blurb-re

该数据集包含两个配置：chemprot和ddi。chemprot配置用于化学蛋白质关系抽取任务，包含文本、实体（如化学物质和蛋白质）及其关系标注。ddi配置用于药物相互作用抽取任务，涉及药物实体及其相互作用关系。每个配置包括训练、验证和测试集，特征包括id、文本、实体列表（含文本、标签和角色）和关系列表（含头尾实体索引和标签）。数据集适用于自然语言处理中的信息抽取和关系分类研究。

This dataset includes two configurations: chemprot and ddi. The chemprot configuration is designed for chemical-protein relation extraction tasks, containing text, entities (such as chemicals and proteins), and their relation annotations. The ddi configuration is used for drug-drug interaction extraction tasks, involving drug entities and their interaction relationships. Each configuration comprises train, validation, and test splits, with features including id, text, a list of entities (with text, label, and role), and a list of relations (with head and tail indices and label). The dataset is suitable for information extraction and relation classification research in natural language processing.