Jaehun/infosumm

该数据集是一个用于文本摘要任务的数据集，包含训练集和验证集。每个样本包含以下特征：文本（text）、摘要（summary）、ID（id）、关键词（keywords）、句子数量（num_sentences）、摘要的提取性（extractiveness）和特异性（specificity）。训练集包含4,704,648个样本，验证集包含50,000个样本。

This dataset is designed for text summarization tasks, containing both training and validation sets. Each sample includes the following features: text, summary, ID, keywords, number of sentences (num_sentences), extractiveness of the summary, and specificity. The training set consists of 4,704,648 samples, and the validation set contains 50,000 samples.