This project is to attempt to increase the speed of convergence of AlexNet using the Jaynes Machine distribution.
We run experiment.ipynb to train AlexNet using the original training parameters. Then we find the lognormal parameters. Finally, we attempt different approaches to speed of training using these lognormal parameters.