In this file, I will update as many as deep learning methods i can complish by using Python. Codes are based on package mxnet.
It shows a possible relization of DFN or MLP. It shows a possibility of retrain model by half learning rate if loss function return NAN.
I'll update RNN and LSTM later.