issues
search
inspire-group
/
hydra
Code and checkpoints of compressed networks for the paper titled "HYDRA: Pruning Adversarially Robust Neural Networks" (NeurIPS 2020) (https://arxiv.org/abs/2002.10509).
http://vsehwag.github.io/hydra
90
stars
24
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Troubles loading checkpoint
#9
giorgiopiras
closed
1 year ago
1
.tar files from dropbox won't extract
#8
giorgiopiras
closed
1 year ago
1
raise exception in the case of unsupported layer types
#7
NormalUhr
closed
1 year ago
0
The bounded model converter will disregard the layer type Dense and make the cifar_model_resnet incorrect.
#6
NormalUhr
closed
1 year ago
0
cannot reach 57.2 on the cifar10 dataset int the pretraining
#5
MVPR-Group
closed
1 year ago
1
Bug fix: CIFAR100 wrongly use CIFAR10 as test dataset
#4
Cyfine
closed
2 years ago
0
Runtime Error
#3
3droot
closed
2 years ago
0
eval error: Pretrained model on ImageNet
#2
byerose
closed
2 years ago
1
Regarding HYDRA for filter pruning
#1
zjysteven
closed
3 years ago
2