文檔中包含網(wǎng)盤的地址，數(shù)據(jù)共319M NLP方向文本摘要，文本分類，等方向可采納！ The LCSTS dataset includes two parts: /DATA: 1. PART I: is the main contents of LCSTS that contains 2,400,591 (short text, summary) pairs. It can be used to train supervised learning models for summary generation. 2. PART II: contains 10,666 human labled (s