Exercise:Sparse Autoencoder

From Ufldl

Jump to: navigation, search
Line 15: Line 15:
Specifically, in this exercise you will implement a sparse autoencoder,  
Specifically, in this exercise you will implement a sparse autoencoder,  
trained with 8×8 image patches using the L-BFGS optimization algorithm.
trained with 8×8 image patches using the L-BFGS optimization algorithm.
 +
 +
'''A note on the software:''' The provided .zip file includes a subdirectory
 +
<tt>minFunc</tt> with 3rd party software implementing L-BFGS, that
 +
is licensed under a Creative Commons, Attribute, Non-Commercial license. 
 +
If you need to use this software for commercial purposes, you can
 +
download and use a different function (fminlbfgs) that can serve the same
 +
purpose, but runs ~3x slower for this exercise (and thus is less recommended).
 +
You can read more about this in the [[Fminlbfgs_Details]] page.
 +
 +
===Step 1: Generate training set===
===Step 1: Generate training set===

Revision as of 00:55, 22 April 2011

Personal tools