search

nmndeep / revisiting-at

[NeurIPS 2023] Code for the paper "Revisiting Adversarial Training for ImageNet: Architectures, Training and Generalization across Threat Models"
37 stars 3 forks source link

Pretrained weights for models mentioned in Table 2 #3

Closed HashmatShadab closed 1 year ago

HashmatShadab commented 1 year ago

Can you please provide weights for other models mentioned in Table 2 such as ViT-S , ConvNeXt-T. ViT-B and ConvNeXt-B(from Table 3)?

nmndeep commented 1 year ago

Hi, you can find these models at the following URL: https://nc.mlcloud.uni-tuebingen.de/index.php/s/XLLnoCnJxp74Zqn

HashmatShadab commented 1 year ago

Thank you for sharing!

HashmatShadab commented 1 year ago

Are the clean pre-trained weights available for the ConvStem models?

nmndeep commented 1 year ago

Hi, the weight-links in the README table should also have clean CvSt checkpoints. Apart from these no other clean-CvSt models are available. Hope this helps.