Skip to content

random-guest/Hessian-Free-Gradient-Flow-Algorithm

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

30 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Quick Intro:

This folder contains all the functions and scripts used to implement the research paper titled "A Hessian-Free Gradient Flow (HFGF) method for the optimization of deep learning neural networks".

How to run the code?

In order to test the code, and obtain the result which is Table 2 in the main research paper, which can be accessed below, download this folder, add its path to the Matlab, and simply run Test_HFGF script.

Note:

Due to the limited access to a high computational power machine, the high dimensional functions were not tested. Feel free to test them if possible.

What is happening underneath the hood?

Test_HFGF script will call the main code of the paper named main_HFGF and will send 4 different testing functions and return the information needed to construct the table. Inside the main_HFGF, there is a call to Armijo function to obtain the step size alpha.

Reference:

https://www.sciencedirect.com/science/article/abs/pii/S0098135420303562

Contact information:

atrashabdulkarim@gmail.com

About

An optimization algorithm for deep neural networks.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published