locuslab/convex_adversarial

Issues

• Support in dual_layers for MaxPool2d

  #35 opened 3 years ago by kkarrancsu
  0

• For version 2.0 i.e, original paper 2D.ipynb gives error

  #36 opened a year ago by aknirala
  0

• Computation time

  #30 opened 5 years ago by sungyoon-lee
  7

• PGD attack error

  #34 opened 4 years ago by jhss
  1

• Is it possible to define DualConvTranspose2d?

  #33 opened 4 years ago by jhss
  1

• Lower bound on the robust radius

  #32 opened 4 years ago by choltz95
  1

• Unable to load the .pth files?

  #10 opened 5 years ago by VSehwag
  2

• epsilon in CIFAR10

  #31 opened 5 years ago by sungyoon-lee
  2

• Loading the state_dict for a model

  #29 opened 5 years ago by and-lap
  1

• Which model should I use?

  #27 opened 5 years ago by dongyp13
  1

• can the pre-trained model get the 79.25%?

  #28 opened 5 years ago by Equationliu
  0

• how to define a reshape layer?

  #26 opened 5 years ago by Equationliu
  1

• Question: Can this work for non-one hot labels?

  #25 opened 6 years ago by shaform
  0

• models_scaled/cifar_small_2px.pth is missing

  #22 opened 6 years ago by vtjeng
  0

• scaled_models/mnist_large_0_1.pth appears to be for network of "small" architecture

  #21 opened 6 years ago by vtjeng
  0

• Constructing a DualNetwork instance with an a single-element np.ndarray `epsilon` value fails.

  #24 opened 6 years ago by gauravmm
  0

• Feature Request: Support PyTorch 1.0.0

  #23 opened 6 years ago by gauravmm
  0

• cannot import name robust_loss_batch

  #18 opened 6 years ago by hope-yao
  2

• CIFAR pretrained model?

  #19 opened 6 years ago by chenhongge
  1

• CIFAR Pretrained model?

  #20 opened 6 years ago by chenhongge
  0

• GPU Out of memory when switching from regular to robust training

  #17 opened 6 years ago by stefanwebb
  2

• Get stuck in training cifar10 on pytorch 0.4.0

  #16 opened 6 years ago by cxmscb
  3

• it is possible to change conv2d's padding to zero?

  #15 opened 6 years ago by cxmscb
  4

• Get stuck in fashion_mnist.py on pytorch 0.4.0

  #14 opened 6 years ago by cxmscb
  0

• Question: Support for non-straight network shapes?

  #12 opened 6 years ago by progirep
  7

• Add None-cases for nonapplicable layers for layers

  #13 opened 6 years ago by riceric22
  0

• DualReLU.affine uses CUDA specific code

  #7 opened 6 years ago by riceric22
  0

• Add warning if provided network is not a sequential module

  #8 opened 6 years ago by riceric22
  0

• Using `nn.Sequential` as the identity function fails with exception.

  #9 opened 6 years ago by gauravmm
  0

• example errors

  #6 opened 7 years ago by andrewcz
  5

• exact settings to reproduce 5.82% robust error for MNIST

  #5 opened 7 years ago by gwding
  5

• Port code to PyTorch 0.4

  #4 opened 7 years ago by riceric22
  1

• `.pth` files not available

  #3 opened 7 years ago by vtjeng
  1

• Training using examples/mnist.py appears to get stuck - with default configuration and various other configurations

  #2 opened 7 years ago by kaixiao
  2

• `MaskedSelect can't differentiate the mask` with torch 0.3.0 and 0.3.1

  #1 opened 7 years ago by vtjeng
  2