To replicate results on a cluster with slurm, first adjust single.bash and single_by_M.bash. Then, run:

mkdir ~/Res-Synth/
mkdir ~/Res-Synth-M/
sbatch main.bash
sbatch main_by_d.bash
sbatch main_by_lambda.bash
python src/figures_res_synth.py
python src/figures_res_synth_by_M.py

You can run local_main.bash etc. for testing code locally.

To replicate results, first download Sushi Dataset (sushi3-2016.zip) and extract to ~/Data/. Then, run:

mkdir ~/Res-Sushi/
python src/data.py
python sushi.py 1
python src/figures_res_sushi.py

Please consider citing our paper if you use this repository.