MadryLab cifar10_challenge issues

MadryLab / cifar10_challenge

A challenge to explore adversarial robustness of neural networks on CIFAR10.

MIT License

488 stars 133 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

pretrained model link expired

#30 hank08tw closed 3 years ago
1
Submitting my result to white CIFAR-10 leaderboards

#29 liuye6666 closed 4 years ago
1
lack of Sign function

#28 Sumching closed 4 years ago
0
The config of training robust model for CIFAR10.

#27 ylsung closed 4 years ago
1
pytorch definition of the model

#26 kartikgupta-at-anu closed 4 years ago
1
clean accuracy on adversarially trained cifar10 resnet 18

#25 wahrheit-git closed 4 years ago
1
Naturally trained network gives 78% test accuracy on pgd attack.

#24 baytasin closed 4 years ago
1
About the network architecture

#23 dongyp13 closed 4 years ago
1
Making adversarial examples during training

#22 symoon11 closed 4 years ago
1
Overflow when random_restart is false

#21 hanboa closed 4 years ago
1
Number of trainable parameters

#20 JonathanCMitchell closed 4 years ago
1
When generating uniform noise in random start, floating point number will cause invalid pixel value.

#19 Line290 closed 4 years ago
1
PGD steps along the sign of the gradient

#18 SohamTamba closed 5 years ago
1
About the convergence of training.

#17 anonymous530 closed 5 years ago
2
Matching training statistics

#16 JonathanCMitchell closed 5 years ago
2
Dataset normalization

#15 TimurIbrayev closed 5 years ago
2
Same accuracy logged twice

#14 JonathanCMitchell closed 5 years ago
1
Base network questions and implementation.

#13 JonathanCMitchell closed 5 years ago
3
Googlenet with owndata

#12 rajasekharponakala closed 5 years ago
1
Image Channels

#11 rajasekharponakala closed 5 years ago
1
About the accuracy of adversarial examples

#10 lith0613 closed 5 years ago
1
about the loss in the pgd_attack

#9 lith0613 closed 5 years ago
1
White-box result of madry_lab_challenges in examples of cleverhans.

#8 lepangdan closed 5 years ago
1
config for 20 iteration pgd that gives 47.04% accuracy.

#7 msingh27 closed 5 years ago
1
Questions about recreating paper results

#6 inkawhich closed 5 years ago
4
How to determine "best" model

#5 inkawhich closed 5 years ago
4
Image out of valid range for the first iteration of PGD attack

#4 vipinpillai closed 6 years ago
2
cifar_input.py Function get_next_batch() has a small bug.

#3 Mrxiaoyuer closed 6 years ago
1
Correct architecture description

#2 wh0 closed 7 years ago
1
pgd_attack: Reconcile C&W loss with @carlini's implementation

#1 wh0 closed 7 years ago
1