Skip to content

Latest commit

 

History

History
14 lines (11 loc) · 909 Bytes

README.md

File metadata and controls

14 lines (11 loc) · 909 Bytes

Backpropagation-C

Designer: Junbo Zhao, Wuhan University, Working in Tsinghua National lab of intelligent images and documents processing.
Email: [email protected] +86-18672365683

Introduction:
This package implements the famous backpropagation algorithm. You can use it when training a Neural Network, or an Autoencoder.

Relationship with Deep Learning
Likewise, you can increase the number of layers to implement a deeper structure, to follow the trend of "Deep Learning". But bear in mind that DL has lots of beautiful tricks, and merely making it deeper will not yield good results!!

Important
I just write bp.cpp as an example. In practice, you should use your own training data and reset the macros.
Also you could modify it by simply repeating some main steps, if you want to increase the number of layers.