Hyper Parameter Optimization System

The project’s purpose is to facilitate the process of training and optimizing models. It should easily allow the developer to create varied algorithms for experimenting. Then these experiments should be easily visualized and stored. The optimization process should include different algorithms such as Bayesian optimization, grid search and random search. The algorithms should also support parallel training for lowering the time spent training all experiments. The visualization should be responsive and updated in real time so the developer can easily decide to stop or restart the optimization process.