Commits

History for

Commits on Jan 6, 2014

Fix typo and fix reference to old class that was renamed
nouiz
committed
00961d2

Commits on Dec 30, 2013

Change references to SigmoidalLayer to HiddenLayer
zacstewart
committed
a376f41

Commits on Nov 25, 2013

Always load the data from the data folder.
nouiz
committed
c475870

Commits on Mar 27, 2013

Also report best_iter + 1 as best iteration
lamblin
committed
2bd0018
Fix off-by-one error in numbering of epochs in most tutorials
lamblin
committed
2dd943a

Commits on Jan 31, 2013

Use list of pairs for updates, not dict, to remove warning
lamblin
committed
0440226

Commits on Oct 31, 2012

make the number of hidden unit for mlp parametrisable.
nouiz
committed
eb315c6
small formating change
nouiz
committed
7ef0549

Commits on Sep 6, 2012

use borrow=True when making shared variable to make the example work with 32 bits python.
nouiz
committed
8596e69

Commits on Jun 16, 2012

update the best_iter variable in the mlp.py example.
nouiz
committed
bb21f3e

Commits on Mar 20, 2012

indent.
nouiz
committed
362a8c0
pep8
nouiz
committed
da46247
remove trailing white space.
nouiz
committed
69d7c41

Commits on Dec 8, 2011

mlp.HiddenLayer now accepts parameters W and b by reference
gdesjardins
committed
5bc54bd

Commits on Jul 17, 2011

Use numpy.inf instead of float('inf').
lamblin
committed
230e17f

Commits on Mar 15, 2011

Remove use of .value in mlp
lamblin
committed
46fc8c9

Commits on Jul 22, 2010

fix syntax
nouiz
committed
d97f381

Commits on Jul 19, 2010

added comment.
nouiz
committed
d3ab267
Merge git://github.com/fidlej/DeepLearningTutorials
nouiz
committed
d7ea6c2

Commits on Jul 16, 2010

fix type and doc formating.
nouiz
committed
e371d52
Unified the W_values initialization.
fidlej
committed
8469cc9

Commits on Apr 1, 2010

removed the expected time from the tutorial code. Now into test.py:speed
nouiz
committed
a8ee2e6

Commits on Mar 26, 2010

merging
Razvan Pascanu
committed
a59a2bf

Commits on Mar 19, 2010

moved the timing into the code to time only the exec time.
nouiz
committed
223e870

Commits on Mar 16, 2010

first draft DBN + sqrt(6/(n_in+n_out)) + different initializations for tanh/sigmoid
Razvan Pascanu
committed
1172608

Commits on Feb 19, 2010

I merged my version 2 into the official deep learning tutorials
Razvan Pascanu
committed
d71dce9

Commits on Feb 12, 2010

a few typos + a few updates to rbm
Razvan Pascanu
committed
cffce58

Commits on Feb 10, 2010

added links to the stacked denoising autoencoder tutorial
Razvan Pascanu
committed
dffd751

Commits on Feb 8, 2010

working version of SdA.py
Razvan Pascanu
committed
c8fc56d

Commits on Feb 5, 2010

Using index instead of offset ; some bug fixes for the SdA.py; partially finished SdA tutorial ( I got stucked in trying to fix the code and didn't manage to finish everything )
Razvan Pascanu
committed
9d2f5d3

Commits on Feb 3, 2010

I have made all tutorials (code and documentation) to use shared variable to store dataset as James suggested. I also changed the validation_frequency to Yoshua's suggestion. One should still need…
pascanur
committed
a074b32

Commits on Feb 2, 2010

Changing to the new way of storing the dataset (using shared variables)
pascanur
committed
ffa5bf9

Commits on Jan 28, 2010

Fixed signed bug in weight initialization
pascanur
committed
4256877

Commits on Jan 22, 2010

latest updates to convolutional code
gdesjardins
committed
e0e5e5d

Commits on Jan 21, 2010

SdA.py implements a stacked de-noising autoencoder in the tutorial style. I think DBN/SdA are a bit different from the rest, so if anybody has the time to read through it and give some feedback I w…
pascanur
committed
76f20c3