Quest-AI/quest-corruption-393k

Quest-corruption 393K数据集是一个用于训练CLM（Character-Level Model）恢复器的数据集。它包含文本内容、唯一标识符、token数量以及数据是否损坏的信息。数据集分为训练集和测试集，其中训练集大小约为3.81GB，包含385,024个样本；测试集大小约为81.11MB，包含8,192个样本。总下载大小约为3.22GB，整个数据集大小约为3.89GB。

The Quest-corruption 393K dataset is designed for training CLM (Character-Level Model) restorers. It includes information such as text content, unique identifiers, token counts, and whether the data is corrupted. The dataset is split into a training set and a test set, with the training set being approximately 3.81GB in size and containing 385,024 samples, while the test set is approximately 81.11MB in size and contains 8,192 samples. The total download size is about 3.22GB, and the entire dataset is about 3.89GB in size.