fwd4xl/enhanced_image_captions

该数据集包含图片和对应的描述（caption），每个样本都有唯一的ID，图片的宽高信息以及宽高比。数据集分为两个部分：h2w_1.0和h2w_0.8，分别包含175,757和19,425个样本。数据集的特征包括：ID（整数型）、图片（图片类型）、描述（字符串型）、宽度（整数型）、高度（整数型）、宽高比（浮点型）和增强描述（字符串型）。

The dataset consists of images and their corresponding captions. Each sample has a unique ID, image width and height information, and aspect ratio. The dataset is split into two parts: h2w_1.0 and h2w_0.8, containing 175,757 and 19,425 samples respectively. The features of the dataset include: ID (int64), image (image type), caption (string), width (int64), height (int64), aspect ratio (float64), and enhanced caption (string).