auto_LiRPA
auto_LiRPA copied to clipboard
auto_LiRPA: An Automatic Linear Relaxation based Perturbation Analysis Library for Neural Networks and General Computational Graphs
Hi and thank you for this great library ! I was wondering it auto Lirpa could be applied on other domains than Lp balls on minibatches ? For example hyper-rectangles...
Hello, Is it possible to compute bounds for a batch of inputs with different epsilon bounds? I discovered it could be done for L_inf norm, but non-trivial for other norms....
Hi, Thank you for making this library available! I was trying to reproduce the results from your paper on the CIFAR-10 dataset. Specifically, I followed the instructions available [here](https://auto-lirpa.readthedocs.io/en/latest/paper.html) and...
@huanzhang12 However, such obtained bounds will not be the same as in CROWN method, won't they? I wanted to look at intermediate bounds from particular methods. Is it possible to...
conv_3d
Hello, I am wondering if the library provides access to bounds of models like C3D? Thanks, Zoeyu
Hi 😄 I'm trying to run auto_LiRPA on a transformer model for token classification (named entity recognition) from the Hugging Face library, but I have run into some problems. Here's...
Dear auto_LiRPA Team, thanks for putting together this library! I would like to use it to verify a network with in-built normalization. Here is an example with an ACAS Xu-like...
I am experiencing an issue when attempting to apply auto_LiRPA to a standard Vision Transformer architecture inspired on the lucidrains implementation available at: https://github.com/lucidrains/vit-pytorch/blob/main/vit_pytorch/vit.py The following assertion error arises when...
I want to use auto_LiRPA to compute the boundary of a neural network with GELU activation function. Here is my code: ```import torch.nn as nn import torch from auto_LiRPA import...
**Describe the bug** I am trying to bound a `clamp` operation. Since using `torch.clamp` directly produces an error stating that `Cast` is an unsupported operation, I instead tried `torch.minimum(torch.maximum(x, mins),...