About The Project

Train and test a multi layer network.

What I learned from this project

• Implement a deep network from scratch
• Use of python libraries like numpy, matplotlib and tensorflow
• Better understanding of the theory of gradient descent and back propagation
• Implement a k-layer deep network from scratch, with batch noramlization and scale and shift.
• Better understanding of initialization of the parameter of network (Xavier/He/Glorot Initialization)
• Coarse Search of hyperparameter like lambda

Getting Started

1. Open anaconda command prompt, cd to the desired lab and run:

   jupyter notebook