search

MadryLab / robustness

A library for experimenting with, training and evaluating neural networks, with a focus on adversarial robustness.
MIT License
905 stars 181 forks source link

Image preprocessing before inference? #74

Closed MalharJ closed 4 years ago

MalharJ commented 4 years ago

For normal PyTorch models we'd usually perform some sort of image preprocessing before inference (https://pytorch.org/docs/stable/torchvision/models.html), with a mean and a standard deviation. Do we do the same for the robust pre-trained models in this library?

andrewilyas commented 4 years ago

Yep! Check out the AttackerModel class, normalization is implemented there.

andrewilyas commented 4 years ago

Closing this now, let me know if you have any questions!