Commits

Breadcrumbs

History for

onmaster

User selector

All users

All time

Commit History

Commits on Sep 1, 2010

• *internal

  

  dmitry

  committed
  3eadd27

  Copy full SHA for 3eadd27

• *trying dA with absolute value reconstruction error

  

  dmitry

  committed
  4379dbd

  Copy full SHA for 4379dbd

Commits on Jul 16, 2010

• fix type and doc formating.

  nouiz

  committed
  e371d52

  Copy full SHA for e371d52

Commits on Apr 2, 2010

• new results for SdA

  

  Razvan Pascanu

  committed
  ee64525

  Copy full SHA for ee64525

Commits on Apr 1, 2010

• removed the expected time from the tutorial code. Now into test.py:speed

  nouiz

  committed
  a8ee2e6

  Copy full SHA for a8ee2e6

Commits on Mar 26, 2010

• merging

  

  Razvan Pascanu

  committed
  a59a2bf

  Copy full SHA for a59a2bf

Commits on Mar 19, 2010

• moved the timing into the code to time only the exec time.

  nouiz

  committed
  223e870

  Copy full SHA for 223e870

Commits on Mar 16, 2010

• first draft DBN + sqrt(6/(n_in+n_out)) + different initializations for tanh/sigmoid

  

  Razvan Pascanu

  committed
  1172608

  Copy full SHA for 1172608

Commits on Feb 24, 2010

• I rerun the Sda tests as a sanity check to all the changes that have been made .. and updated the results. I did not get the time though to look into running the SdA with parameters such that you …

  Show description for 3a3fb4d
  

  Razvan Pascanu

  committed
  3a3fb4d

  Copy full SHA for 3a3fb4d

Commits on Feb 19, 2010

• I merged my version 2 into the official deep learning tutorials

  

  Razvan Pascanu

  committed
  d71dce9

  Copy full SHA for d71dce9

Commits on Feb 15, 2010

• fix syntax error.

  nouiz

  committed
  356d814

  Copy full SHA for 356d814

Commits on Feb 12, 2010

• a few typos + a few updates to rbm

  

  Razvan Pascanu

  committed
  cffce58

  Copy full SHA for cffce58

Commits on Feb 11, 2010

• use floatX dtype for bias - fixed

  

  Razvan Pascanu

  committed
  878f952

  Copy full SHA for 878f952

• removed numpy_rng which I wasn't using anyhow

  

  Razvan Pascanu

  committed
  f2d3749

  Copy full SHA for f2d3749

Commits on Feb 10, 2010

• SdA_v2 mainstream .. deleted depricated SdA version

  

  Razvan Pascanu

  committed
  7358e0e

  Copy full SHA for 7358e0e

Commits on Feb 8, 2010

• added corruption level as a hyper-parameter

  

  Razvan Pascanu

  committed
  146fda1

  Copy full SHA for 146fda1

• Merge branch 'master' of git@github.com:lisa-lab/DeepLearningTutorials

  

  Razvan Pascanu

  committed
  5832e1d

  Copy full SHA for 5832e1d

• working version of SdA.py

  

  Razvan Pascanu

  committed
  c8fc56d

  Copy full SHA for c8fc56d

Commits on Feb 7, 2010

• added comment

  

  Yoshua Bengio

  committed
  6500270

  Copy full SHA for 6500270

Commits on Feb 5, 2010

• Merge branch 'master' of git@github.com:lisa-lab/DeepLearningTutorials

  pascanur

  committed
  69c8ee9

  Copy full SHA for 69c8ee9

• Merge branch 'master' of git@github.com:lisa-lab/DeepLearningTutorials

  pascanur

  committed
  f84b517

  Copy full SHA for f84b517

• merge

  pascanur

  committed
  7ffaf89

  Copy full SHA for 7ffaf89

• merge

  

  Razvan Pascanu

  committed
  0301c94

  Copy full SHA for 0301c94

• A more readable/almost complete version of the SdA tutorial.. still working on a possible bug in the SdA.py

  

  Razvan Pascanu

  committed
  d6c0574

  Copy full SHA for d6c0574

• Changed variable name from n_iter to n_epochs for clarity

  

  Razvan Pascanu

  committed
  d2f54e0

  Copy full SHA for d2f54e0

• 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

  Copy full SHA for 9d2f5d3

Commits on Feb 4, 2010

• hyperparameters of SdA

  pascanur

  committed
  cc97220

  Copy full SHA for cc97220

Commits on Feb 3, 2010

• Updated SdA code to be ready for publishing .. I also chose the version of SdA with variable number of layers to be the official tutorial

  pascanur

  committed
  0b302a0

  Copy full SHA for 0b302a0

Commits on Jan 27, 2010

• A small bug in both version of stacked denoising autoencoders fixed

  pascanur

  committed
  d068e5d

  Copy full SHA for d068e5d

Commits on Jan 22, 2010

• Fixed a few bugs and defined a SdA with variable size

  pascanur

  committed
  c9e99a6

  Copy full SHA for c9e99a6

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…

  Show description for 76f20c3
  pascanur

  committed
  76f20c3

  Copy full SHA for 76f20c3

Loading