amnaalmgly / axiomaticdeepnets

AxiomaticDeepNets

paper Reference: https://arxiv.org/abs/1703.01365

I was able to interperate images data classification (on pretrained googlenet )as well as IMDB data sentiment Analysis(on biderctional LSTM model) and visualize the results

Major modules implemented in the code

• Interpolation
• Gradient
• Integrated Gradient
• Visualization

How to use code

Process your Experiment(the image / text ) you want to explain as follows:

• upload your image in the /Images folder
• edit your config file with the path of the image ,result to be saved, pretrained model path and imagenet classes.txt in case of images
• run the main.py file
• results are saved in the results folder representing visualizaion of : baseline , original image , gradient and Integrated Image, Overlay the image with IG

Clone the repository

git clone https://github.com/[username]/AxiomaticDeepNets.git

Setup a new environment using requirements.txt in repo

pip3 install -r requirements.txt 

Setup configuration in config.py file

go to src > config.py

Run python main.py with command-line arguments or with edited config file

python main.py 

TODO

1. Improve documentation