MUSIC
收藏arXiv2025-09-30 收录
下载链接:
https://github.com/misonuma/strsum
下载链接
链接失效反馈官方服务:
资源简介:
该数据集包含了来自亚马逊的在线评论及其一句话摘要,总计1.428亿条评论。这些评论涵盖了玩具与游戏、运动与户外以及电影与电视等多个领域。在训练过程中,所使用的评论包含10句或更多句子,而验证和评估环节则采用了含有5句或更多句子的评论。该数据集的规模达到了1.428亿条评论,任务是对评论进行抽象式摘要。
This dataset contains 142.8 million Amazon online reviews paired with their corresponding one-sentence summaries. These reviews cover a wide range of domains including Toys & Games, Sports & Outdoors, Movies & Television, and others. For the training stage, reviews with 10 or more sentences are utilized, while reviews containing at least 5 sentences are employed for validation and evaluation phases. The core task of this dataset is abstractive summarization for user reviews.
提供机构:
Amazon



