The problem of experimental non-convergence

[munan-19]

I'd like to ask why it doesn't converge when I run it with the resnet18 model when reproducing this code?

[munan-19]

Hi, I'd appreciate it if you could answer my question.

[Alexkael]

Hi, could you provide the details? I am not sure whether your implementation of ResNet-18 is suitable with the script, but I am happy to provide my script of ResNet-18 later.

[Alexkael]

Hi, I'd appreciate it if you could answer my question.

Please find train_res18_our.py in https://github.com/Alexkael/Randomized-Adversarial-Training/tree/main/AWP_first_second, if you have further questions, please feel free to ask me.

[munan-19]

嗨，如果您能回答我的问题，我将不胜感激。

请在 https://github.com/Alexkael/Randomized-Adversarial-Training/tree/main/AWP_first_second 找到train_res18_our.py，如果您还有其他问题，请随时问我。

Thank you for your answers and for the help you provided