bigdata-pw/fashion-150k

Fashion-150k数据集包含151,748个产品，587,824张图片，涉及1,186位设计师。数据格式为WebDataset，每个shard包含1k个数据点，大约0.4GB。每个数据点包括图片和JSON文件，JSON文件中包含产品的性别、名称、类别、设计师和价格等信息。数据集支持英语、法语、中文、日语和韩语。

The Fashion-150k dataset includes product images and metadata, suitable for image classification and image-to-text tasks. It contains 151,748 products, 587,824 images, and 1,186 designers, supporting English, French, Chinese, Japanese, and Korean languages. Each product record includes fields such as gender, name, categories, designer, and price, with multilingual translations provided for name and designer fields.