The code must be run from main directory of the repository, not from this directory. For example, if your working directory is mnist, first go up one level

cd ..

Then run the training process using this command:

python mnist/mnist-full-run.py --gpu <choose your GPU number> --names <model-name> --log-file <model-log-file-name>

For example, if you want to run the experiment with NesterovNODE with GPU 0 and let the training log file name be nesterovnode, then the command is as follows:

python mnist/mnist-full-run.py --gpu 0 --names nesterovnode --log-file nesterovnode

Modify this command if you want to run with other models.

The code for plotting the results is in visualization/MNIST_visualization.ipynb. If the training process of the model is done successfully, follow the cells in this .ipynb file to reproduce the plots.