CrowdMind/score-4-fineweb-edu

这个数据集名为score-4-fineweb-edu，它包含了aynetdia/fineweb-edu-score-4-dedup数据集中的text列，并以Parquet格式导出。该数据集仅包含文本数据，没有其他元信息如URL或分数。它是为那些期望纯文本例子的训练数据加载器优化的。数据集的分片是基于字符数的，每个分片大约包含2500万个字符，并且整个数据集被分为18个分片。数据集的行组大小为1024，创建于2025年11月9日。

This dataset, named score-4-fineweb-edu, includes only the text column from the aynetdia/fineweb-edu-score-4-dedup dataset, exported in Parquet format. It contains text data only, without any metadata such as URLs or scores. It is optimized for training data loaders that expect text-only examples. The dataset is sharded based on character count, with each shard containing approximately 25 million characters, and the entire dataset is divided into 18 shards. The dataset has a row group size of 1024 and was created on November 9, 2025.