ljnlonoljpiljm/laion-2b-en-unsafe-quarter-four-download-hd

该数据集包含多个字段，如URL、文本内容、图片宽度、图片高度、相似度、哈希值等。数据集被划分为训练集，其中包含282,285个样本，总大小约为34.34GB。数据集还提供了默认配置，包括训练集数据文件的路径。

The dataset includes multiple fields such as URL, text content, image width, image height, similarity, hash, etc. The dataset is split into a training set, which contains 282,285 samples and has a total size of approximately 34.34GB. The dataset also provides a default configuration, including the path to the training set data files.