NewEden/Big-qwen-OWA-Conv

该数据集包含三个特征：tokens（整数列表）、mask（布尔列表）、labels（整数列表）。数据集仅包含训练集，共有91050个示例，数据大小为3.54GB。数据集的下载大小为536MB，并提供了一个默认配置文件，其中包含了训练数据文件的路径信息。The dataset includes three features: tokens (list of integers), mask (list of booleans), and labels (list of integers). It contains only the training set with a total of 91,050 examples, with a dataset size of 3.54GB. The download size of the dataset is 536MB, and a default configuration file is provided, which includes the path information for the training data files.