Skip to content

Latest commit

 

History

History
33 lines (26 loc) · 1.12 KB

INSTALL.md

File metadata and controls

33 lines (26 loc) · 1.12 KB
Raw

INSTALL

The following steps should be sufficient to get these attacks up and running on most systems running Python 3.7.3+.

  1. Download a suitable version of PyTorch for your environment.

    • This project originally ran with torch==1.5.0 and torchvision==0.6.0, but other versions are likely to work as well.
    • It is also highly recommended that you use GPUs to execute the evaluation scripts.

  2. Run the requirements.txt file:

numpy       ~=1.16.2
pandas      ~=0.24.2
tqdm        ~=4.31.1
matplotlib  ~=3.0.3
scipy       ~=1.2.1
seaborn     ~=0.9.0

Note: these are the most recent versions of each library used, lower versions may be acceptable as well. If you have other versions Python, the ~= operator should download the best compatible version of each package for you.

pip install -r requirements.txt
  1. Viewing results in Jupyter Notebooks

The results are aggregated and visualized in a jupyter notebook, which can be viewed directly in GitHub or perused locally:

# install
pip install jupyter

# start notebook in working directory
jupyter notebook