aamir-mustafa pcl-adversarial-defense issues - Githubissues

aamir-mustafa / pcl-adversarial-defense

Adversarial Defense by Restricting the Hidden Space of Deep Neural Networks, in ICCV 2019

59 stars 14 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Mapping function in code

#18 pankhurivanjani opened 2 years ago
0
Models used for MNIST

#17 ashwath98 opened 2 years ago
0
Lpc loss

#16 Jeevi10 opened 3 years ago
0
The performance of provided "robust_model.pth.tar" is not ideal

#15 Harry-Up opened 4 years ago
2
Missing CW attack

#14 shairoz closed 4 years ago
4
Do we need to load softmax_train.py trained model and then load the pretrained model in pcl_training.py

#13 machanic closed 4 years ago
1
Clear up confusion with the Proximity and Con_Proximity

#12 rowedenny closed 4 years ago
1
Clearing up some concepts

#11 prateekm08 closed 4 years ago
1
Cant get the reported accuracy by running the training scripts.

#10 Amose-Yao closed 4 years ago
1
Proxity and Conproxity losses share different prototype centers in code

#9 krishnakanthnakka closed 4 years ago
1
Loss weights for different datasets

#8 krishnakanthnakka closed 4 years ago
1
What is "baseline" mean on table 8 & 9

#7 rowedenny closed 4 years ago
1
Softmax model not converging

#6 krishnakanthnakka closed 4 years ago
2
CIFAR10_PCL_AdvTrain_PGD.pth.tar and CIFAR10_PCL_AdvTrain_FGSM.pth.tar ?

#5 AliShahin closed 4 years ago
1
Wrong normalization

#4 Line290 closed 4 years ago
1
loss function in eq5 goes negative infinity?

#3 rowedenny closed 4 years ago
1
black box attack

#2 wahrheit-git closed 4 years ago
4
Update README.md

#1 salman-h-khan closed 5 years ago
0