ardauzunoglu/dclm_random200m

该数据集包含197,212个训练样本，总大小约为12.1GB，下载大小约为734.7MB。每个样本包含以下字段：sample_index（样本索引，int64类型）、scanned_index（扫描索引，int64类型）、text（文本内容，字符串类型）、source_dataset（来源数据集名称，字符串类型）和url（来源链接，字符串类型）。数据仅提供训练分割，文件路径为data/train-*。This dataset contains 197,212 training samples with a total size of approximately 12.1GB and a download size of approximately 734.7MB. Each sample includes the following fields: sample_index (int64), scanned_index (int64), text (string), source_dataset (string), and url (string). The data is provided only in a training split, with file paths as data/train-*.