LG-PGN

This dataset is a Chinese dataset, derived from news_2016zh, and has been filtered and segmented by characters. The dataset contains 300,000 training samples, 12,000 validation samples, and 12,000 test samples.本数据集为中文语料数据集，源自news_2016zh，经字符级过滤与切分处理。该数据集包含30万个训练样本、1.2万个验证样本及1.2万个测试样本。