This repository contains the experiments found in the paper: https://arxiv.org/abs/1703.08961
In order to run our experiments you will need at minimum the following python packages: pytorch,opencv,pyscatwave package. The simplest way to install pytorch and opencv is through anaconda. We recommend python 2.7 + anaconda. The pyscatwave package can be found here https://github.com/edouardoyallon/pyscatwave
We provide a pre-trained model similar to the one described in the paper.
To run the trained model of scattering+resnet on imagenet ILSVRC validation set:
Training scripts for imagenet and SLE feature extractor will be added soon
Simply run python main_STL.py script in the STL directory
To run the small sample experiments Example:
python main_small_sample_class_normalized.py --model resnet12_8_scat --save "test" --seed 1 --sampleSize 500 --mul 20