How to Make Deep Learning Models That Don’t Suck (blog.nanonets.com)

Training deep learning models can be tough. They don't work without the right hyperparameters. The interactive blog gives an explanation of the algorithms that can automate the hyperparameter search and has all the code you need to try it out for yourself.

Great article! Might be worth adding that too much dropout can cause your model to never converge, which would be creating a model that "sucks!"