Welcome to the official code repository for our submission: "Neural Additive Image Model: Interpretation through Interpolation". This repository contains the code for all experiments conducted in the paper.
This repository currently serves as the code basis for our submission and is therefore maintained in an anonymized format to comply with review requirements. Upon acceptance of our work, we plan to release the complete code in a more comprehensive package format, making it publicly available for the community.
To install the required dependencies, please run the following commands:
pip install -r requirements.txt
git clone https://github.com/phizaz/diffaeThe folder "Preprocessing" contains all code used for preprocessing of the data. The folder "NAIM" has the nessesary code to train and apply Neural Additive Image Models and "Ablation" contains the code for the comprehensive Ablation study in our paper. The "Notebooks" folder contains experimental code.