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

该数据集包含了URL、文本内容、图像尺寸、相似度、哈希值、不安全概率、水印概率等字段的信息。数据集被划分为训练集，包含超过140万个示例，整个数据集大小超过33吉字节。具体的应用场景和文本内容、图像内容等信息未在README中说明。

The dataset includes information such as URL, text content, image dimensions, similarity, hash value, probability of being unsafe, probability of watermark, etc. The dataset is split into a training set, containing more than 1.4 million examples, and the entire dataset size is over 33 gigabytes. The specific application scenario and the content of the text and images are not described in the README.