Code for the paper "Generalizing to Unseen Domains via Adversarial Data Augmentation", NIPS 2018
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
.gitignore
LICENSE
README.md
download_and_process_mnist.py
download_svhn.sh
exp_configuration
main.py
model.py
run_exp.sh cleanup Sep 13, 2018
trainOps.py
utils.py

README.md

Code for the paper Generalizing to Unseen Domains via Adversarial Data Augmentation

Overview

Files

model.py: to build tf's graph

trainOps.py: to train/test

exp_configuration: config file with the hyperparameters

Prerequisites

Python 2.7, Tensorflow 1.6.0

How it works

To obtain MNIST and SVHN dataset, run

mkdir data
python download_and_process_mnist.py
sh download_svhn.sh

To train the model, run

sh run_exp.sh GPU_IDX

where GPU_IDX is the index of the GPU to be used.