An interactive visualization system designed to help non-experts learn about the Manifestation of Adversarial Examples on Convolutional Neural Networks (CNNs).
For a live demo, visit: https://advvis-cnn.raccooncc.top/
Clone or download this repository:
git clone git@github.com:fireraccoon/AdvVis-CNN.git
# use degit if you don't want to download commit histories
degit fireraccoon/AdvVis-CNNInstall the dependencies:
npm installThen run AdvVis CNN:
npm run devNavigate to localhost:5000. You should see AdvVis CNN running in your broswer :).
To see how to train the CNN, visit the directory ./tiny-vgg/.
To see how to attack for the adversarial samples, visit the directory ./adver-sample/.
AdvVis CNN is based on CNN Explainer, and was created by fireraccoon.
The software is available under the MIT License.
If you have any questions, feel free to open an issue or contact fireraccoon.