许可证: mit
库名称: open_clip
管道标签: 零样本图像分类

[论文] [GitHub]

基于CLIP模型laion/CLIP-ViT-B-16-laion2B-s34B-b88K的鲁棒感知度量

使用FARE（Schlarmann等人 (2024)在ImageNet上以无穷范数和半径4/255进行对抗性微调。

在感知相似性任务NIGHTS上的性能表现：

干净数据     L-无穷, eps=4/255     L2, eps=3  
90.6      71.5                 65.5  

使用方法

model, _, image_processor = open_clip.create_model_and_transforms('hf-hub:chs20/FARE4-ViT-B-16-laion2B-s34B-b88K')  

引用

如果您认为此模型有用，请考虑引用我们的论文：

@inproceedings{croce2024adversarially,  
  title={Adversarially Robust CLIP Models Can Induce Better (Robust) Perceptual Metrics},  
  author={Croce, Francesco and Schlarmann, Christian and Singh, Naman Deep and Hein, Matthias},  
  year={2025},  
  booktitle={{SaTML}}  
}  

@inproceedings{schlarmann2024robustclip,  
    title={Robust CLIP: Unsupervised Adversarial Fine-Tuning of Vision Embeddings for Robust Large Vision-Language Models},   
    author={Schlarmann, Christian and Singh, Naman Deep and Croce, Francesco and Hein, Matthias},  
    year={2024},  
    booktitle={{ICML}}  
}