adversarial training

Follow set-up according to mister_ed. Some names, models, etc. modified, but most lib structure is the same.

E.g. to run PGD-based adversarial training: python -m scripts.advtrain_resnet32_pgd --exp resnet32_pgd --arch resnet32 --verbosity medium

Name		Name	Last commit message	Last commit date
Latest commit History 213 Commits
bin-codes		bin-codes
cifar10		cifar10
custom_lpips		custom_lpips
imagenet		imagenet
mnist		mnist
notebooks		notebooks
scripts		scripts
utils		utils
.DS_Store		.DS_Store
.gitignore		.gitignore
README.md		README.md
__init__.py		__init__.py
adversarial_attacks.py		adversarial_attacks.py
adversarial_evaluation.py		adversarial_evaluation.py
adversarial_perturbations.py		adversarial_perturbations.py
adversarial_training.py		adversarial_training.py
config.json		config.json
config.py		config.py
example_config.json		example_config.json
loss_functions.py		loss_functions.py
main_sandbox.py		main_sandbox.py
prebuilt_attacks.py		prebuilt_attacks.py
prebuilt_loss_functions.py		prebuilt_loss_functions.py
requirements.txt		requirements.txt
runpy.sh		runpy.sh
spatial_transformers.py		spatial_transformers.py
st_sandbox.py		st_sandbox.py

joycex99/adversarial-training