AdoCleanCode/clean_dirty_dac_dac_v1_grpo

该数据集包含噪声类型（noise_type）、序列（sequence）和长度（length）三个字段。噪声类型和序列字段是字符串类型，长度字段是整型。数据集分为一个批次（batch_0000），包含24997个样本，总大小为1.53GB。提供默认配置，数据文件路径为data/batch_0000-*。

The dataset includes three fields: noise_type, sequence, and length. The noise_type and sequence fields are of string type, and the length field is an integer type. The dataset is divided into one batch (batch_0000), containing 24,997 samples, with a total size of 1.53GB. The default configuration is provided, with the data file path set to data/batch_0000-*.