Generate an Adversarial Example

[AhmedHussKhalifa]

Hey @yoshitomo-matsubara,

Can you walk me through how to generate an adversarial example within the dataloader, and what changes would need to be made to ensure that the adversarial samples are properly generated and integrated into the training process?

Thank you in advance.

[yoshitomo-matsubara]

Hi @AhmedHussKhalifa

It sounds like nothing specific for torchdistill, and I would highly suggest you ask such questions at PyTorch Discussion instead.

[yoshitomo-matsubara]

@AhmedHussKhalifa
You've already asked a similar question here

For follow-up questions, please use the same thread rather than starting a new thread, otherwise your question is treated as an answer by the system).

[yoshitomo-matsubara]

@AhmedHussKhalifa

1. Please carefully read and close How to perform identical transform on two different Images? #171 by marking the previous answer as answer
2. Based on my answers at How to perform identical transform on two different Images? #171, re-post the question here (again, as part of this thread, not as an "answer"
3. Delete duplicate "answers" 1 and 2

[AhmedHussKhalifa]

Hey,
Sorry I was not clear.
Let's assume that I have an initialized object for an attack method in ./example/imageclassification.py, and I want to utilize it in BaseDatasetWrapper(). what changes would need to be made to BaseDatasetWrapper()?

[yoshitomo-matsubara]

@AhmedHussKhalifa
I'll stop answering your questions as you don't use this space respectfully, and this is repeated.

How can you

1. close this thread by marking your question as answer (I unmarked it),
2. unmark the answer to your previous question, and
3. copy-paste your question as "answer" in the previous discussion page again, which is exactly what I asked you above not to do
   ?

[AhmedHussKhalifa]

Dear @yoshitomo-matsubara ,

I am writing to apologize for any misunderstandings that may have arisen in our previous interactions. As a new user of the discussion platform, I may have yet to understand the community's etiquette and norms of communication fully.

I understand that my previous questions may have come across as inconsiderate, and I am genuinely sorry for any misunderstanding of your instruction that I may have caused. I appreciate your patience in explaining your expectations to me, and I will do my best to follow them in the future.

Please know that I value your insights and contributions, and I hope to continue learning from you and others in this community.

Please let me know if you accepted my apology. Thank you for taking the time to address my behaviour.