villacu/p50_faces_s1

该数据集包含提示(prompt)、完成(completion)、标签(label)和图像(image)信息。提示和完成部分都包含内容和角色信息，标签为二元判断，图像为图像数据类型。数据集分为训练集、测试集、验证集和配对验证集，各自有不同的数据量和文件大小。默认配置下，提供了各个数据集的文件路径。

The dataset includes prompt, completion, label, and image information. Both prompt and completion sections contain content and role information, the label is a binary judgment, and the image is an image data type. The dataset is split into training set, test set, validation set, and paired validation set, each with different amounts of data and file sizes. Under the default configuration, the file paths for each dataset are provided.