Google's C4 dataset

该数据集是用于训练基于谷歌FNet架构的自动编码器模型的一组新闻故事子集。在经过50个周期的训练后，该模型在这组数据集上达到了96.36%的准确率。此外，一个包含结果的JSON文件可在网上查阅。任务是对新闻故事进行文本生成。This dataset is a subset of news stories used for training autoencoder models based on the Google FNet architecture. After 50 training epochs, the model achieved an accuracy of 96.36% on this dataset. Furthermore, a JSON file containing the results is available online. The task is text generation for news stories.