bobiblazeski / topdown

Top-Down Networks

Top-Down Networks

Code for reproducing the experiments of the paper Top-Down Networks: A coarse-to-fine reimagination of CNNs. In case of any bugs/improvements, please reach Ioannis Lelekas ([email protected]).

Requirements and Dependencies:

We advise Anaconda for quick installation of all requirements and dependencies. Simply use the provided requirements.yml and conda env create -f requirements.yml.

Organization:

This repository is organized as follows:

src/                #source code
  lib/              #helper functions
  models/           #network architectures
  scripts/          #scripts for running the experiments

output/             #output folder; generated after running the code
  adversarial/      #extracted adversarial attacks
  gradcam/          #gradcam heatmaps
  graphs/           #training curves
  history/          #training history (loss, acc, learning curve)
  models/           #model checkpoints
  output/           #generated output from training

notebooks/          #notebooks for visualizing results

Usage:

All scripts for running the experiments are in src/scripts/. Command line inputs are given as comments within the scripts.

Citation:

For citing Top-Down Networks, please use the following:

@inproceedings{lelekas2020,
  title={Top-Down Networks: A coarse-to-fine reimagination of CNNs},
  author={I Lelekas and N Tomen and SL Pintea and JC van Gemert},
  journal={CVPR 2020 Workshop on Deep Vision},
  year={2020}
}

Acknowledgements:

• https://github.com/BIGBALLON/cifar-10-cnn
• https://github.com/bethgelab/foolbox
• https://github.com/eclique/keras-gradcam
• https://github.com/koshian2/PCAColorAugmentation