The sample mainly illustrates the Apache MxNet deep learning framework usage. Like other deep learning framework, MxNet has different data loading strategy, especially for batched data set. There exist 12 sections in the notebook, such as library importation, data loading, hyper-parameter definition, batched data creation, dense MLP definition, etc. See the sections for more details of the learning strategy. The same built-in data set MNIST is used in the example.