Skip to content

Aysegul/torch-nin

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

50 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

torch-NetworkInNetwork

a replication of Network in network in torch. original code : https://github.com/mavenlin/cuda-convnet

Learning rate 0.1 is used in the original code, but it doesn't make the learning happen in this script. Learning rate is set to 2e-3, and the accuracy gets to %87.6 in 500 epochs not %89.6 :(

Another difference is the original code runs on batches of 128, this is 64.

This script works with the last version of optim/sgd.lua Please update

luarocks install optim

About

a replication of Network in network in torch. original code : https://github.com/mavenlin/cuda-convnet

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages