ziyu3141/rf_newtrain_6_5

该数据集包含了图片的质量评估信息，包括美学评分、瑕疵评分、对齐错误评分等。每个样本都包含了多个相关特征，如瑕疵热图、对齐错误热图、对齐错误标记、是否不均匀、首选图片、非首选图片、修改后的图片、未修改的图片ID以及是否首选的标记。数据集分为训练集，提供了相应的字节数和样本数信息。

The dataset consists of image quality assessment information, including aesthetics scores, artifact scores, misalignment scores, and more. Each sample contains multiple related features such as artifact heatmaps, misalignment heatmaps, misalignment token labels, unevenness flags, preferred images, unpreferred images, revised images, unrevised IDs, and preferred flags. The dataset is split into a training set with provided byte size and number of examples.