Fashion-MNIST

Fashion-MNIST是由Zalando研究创建的一个新型图像数据集，包含70,000张28x28像素的灰度图像，展示了10个类别的时尚产品。该数据集旨在作为机器学习算法的基准测试集，特别设计为原始MNIST数据集的直接替代品，共享相同的图像大小和数据格式。数据集内容涵盖男女童中性的时尚产品，通过专业摄影和图像处理技术生成。Fashion-MNIST不仅保持了MNIST的便捷性，还提供了更具挑战性的分类任务，适用于深度学习和机器视觉领域的算法测试与优化。

Fashion-MNIST is a novel image dataset created by Zalando Research, which contains 70,000 grayscale images of 28×28 pixels, showcasing fashion products across 10 categories. This dataset is designed as a benchmark for machine learning algorithms, specifically engineered as a direct drop-in replacement for the original MNIST dataset, sharing identical image dimensions and data formats. The dataset encompasses gender-neutral fashion products for men, women and children, generated via professional photography and image processing techniques. Fashion-MNIST not only retains the convenience of the MNIST dataset but also offers more challenging classification tasks, making it suitable for algorithm testing and optimization in the fields of deep learning and computer vision.