cpsc-5710-final-vit-robustness/patch-attacks-defenses-with-vits

该数据集包含对抗性攻击相关的信息，涉及机器学习模型的攻击和防御类型、原始和预测的标签名称、攻击成功与否的标志，以及输入和输出图像。数据集分为两种配置：detailed（详细）包含具体的攻击实例和图像数据，summary（摘要）提供模型性能的汇总统计如准确率和攻击成功率。

This dataset contains information related to adversarial attacks, including attack and defense types for machine learning models, original and predicted label names, flags for successful attacks, and input and output images. The dataset is divided into two configurations: detailed with specific attack instances and image data, and summary providing aggregated statistics on model performance such as accuracy and attack success rate.