ACL Title and Abstract Dataset

该数据集包含了从ACL选集网络收集的10,874篇论文的标题和摘要配对，用于训练和评估摘要生成模型。该数据集被随机划分为训练集（80%）、验证集（10%）和测试集（10%），规模为10,874组标题和摘要。所涉及的任务是从标题生成摘要。

This dataset includes 10,874 title-abstract pairs collected from the ACL Anthology Network, and is designed for training and evaluating abstractive summarization models. It is randomly partitioned into three subsets: the training set (80%), validation set (10%), and test set (10%), with a total of 10,874 such pairs. The targeted task of this dataset is generating paper abstracts given their titles.