This contains the code necessary to reproduce Figure 1 from the SNN paper. Note that the code uses the biutils package to load the MNIST/CIFAR10 datasets. You can use
pip install git+https://github.com/untom/biutils.git
to easily install the package.
The data for the plot was created by running
./run.py -g 0 -d 08 -a selu -l 1e-5 -e 2000 --dataset mnist
./run.py -g 1 -d 16 -a selu -l 1e-5 -e 2000 --dataset mnist
./run.py -g 2 -d 32 -a selu -l 1e-5 -e 2000 --dataset mnist
./run.py -g 3 -d 08 -a relu --batchnorm -l 1e-5 -e 2000 --dataset mnist
./run.py -g 0 -d 16 -a relu --batchnorm -l 1e-5 -e 2000 --dataset mnist
./run.py -g 1 -d 32 -a relu --batchnorm -l 1e-5 -e 2000 --dataset mnist
./run.py -g 0 -d 08 -a selu -l 1e-5 -e 2000 --dataset cifar10
./run.py -g 1 -d 16 -a selu -l 1e-5 -e 2000 --dataset cifar10
./run.py -g 2 -d 32 -a selu -l 1e-5 -e 2000 --dataset cifar10
./run.py -g 3 -d 08 -a relu --batchnorm -l 1e-5 -e 2000 --dataset cifar10
./run.py -g 0 -d 16 -a relu --batchnorm -l 1e-5 -e 2000 --dataset cifar10
./run.py -g 1 -d 32 -a relu --batchnorm -l 1e-5 -e 2000 --dataset cifar10
The plots where then created using create_plots.ipynb.