auto_LiRPA

auto_LiRPA copied to clipboard

build_gradient_node is not supperted for Sigmoid and Tanh.

2

Thanks for the great work. When I use the Sigmoid activation function. It raises and NotImplementedError: " Function `bound_forward` of `BoundSigmoid(name="/18")` is not supported yet" I think it is because...

Walleclipse

different kernel size and stride in BoundMaxPool

2

Hello, thanks for the great work! I noticed that the BoundMaxPool requires to have equal kernel size and stride https://github.com/Verified-Intelligence/auto_LiRPA/blob/master/auto_LiRPA/operators/pooling.py#L66 Is there any way to remove this constraint? For example,...

xiangchong1

**Describe the bug** `compute_bounds` encounters an error when calling the `get_sparse_C` function through `check_prior_bounds` for certain `nn.Module` configurations. **To Reproduce** 1. Minimum example is provided in [this colab ](https://colab.research.google.com/drive/1RSuThnONpkIb-Dxb7Cl0y6KLTk--Pfy4?usp=sharing) 2....

nrober1122

**Describe the bug** I have found that depending on the configuration of the `nn.Module` being analyzed, the computation time of `compute_bounds` using the `backward` method increases exponentially with network size,...

nrober1122

Errors depending on choice of epsilon-ball for alpha-crown and closed-loop dynamics model

3

**Describe the bug** For the same model, I can compute output bounds successfully for some choices of epsilon-balls, but receive 2 different errors for some other choices of the epsilon-ball....

mfe7

when i use a simple net for testing ``` class simpleNet(nn.Module): def __init__(self) -> None: super(simpleNet,self).__init__() self.linear_input = nn.Linear(64,100) self.relu = nn.ReLU() def forward(self,x): x = self.linear_input(x) x = self.relu(x)...

fxh903

CUDA out-of-memory when running examples/vision/jacobian.py

1

**Describe the bug** When trying the script `examples/vision/jacobian.py` on a 1080Ti I ran into an CUDA out-of-memory error. The log says that I should file a ticket. ``` python jacobian.py...

nhamlv-55

Error with perturbations on some elements in x_L and x_U

1

I extracted a small subnet from ResNet18 as my target network. I specified the upper and lower bounds of the perturbation and used auto_LiRPA to calculate the bounds. However, some...

nnjnjn

Hi, I am trying to use alpha-beta CROWN to verify compositions of trained models and nonlinear functions by wrapping the trained model into a new model with customized `forward` function,...

JianqiangDing

**Describe the bug** I tried to use auto_LiRPA to verify the robustness of my audio processing model. Unfortunately, when I use `model.compute_bounds`, a `RuntimeError: The following operation failed in the...

charnameRed