Note: autograd_hacks.py belongs to https://github.com/cybertronai
Thanks to https://github.com/cybertronai/autograd-hacks for the autograd_hacks.py script
that helps in extracting the gradient information from the networks.

main_mnist.py for experiment 1 uses \tilde{w} from optimal_mnist.npy
main_sm.py for experiment 2 uses \tilde{w} from optimal_sm.npy

The dataset required will be automatically downloaded.
The results will be saved in the result folder.

best() represents DCL+
reverse_best() represents DCL-
vanilla() represents vanilla