Commits · c50f235d96e9f82b1b2c95ce72d4b384a344cba6 · Main Repositories / otb

Sep 08, 2017
- removed the factor of the variance for normalization · c50f235d
  Cédric Traizet authored 7 years ago
  
  c50f235d
- comment added · ac8185b7
  Cédric Traizet authored 7 years ago
  
  ac8185b7
- Neuron type changed to Logistic · c277879e
  Cédric Traizet authored 7 years ago
  
  c277879e
Aug 04, 2017
- bug correction for denoising ae · f3436d08
  Cédric Traizet authored 7 years ago
  
  f3436d08
Jul 27, 2017
- the weight initial value during training can now be set with a parameter · 7b78f3d7
  Cédric Traizet authored 7 years ago
  
  7b78f3d7
- removed irrelevants comments · e9635127
  Cédric Traizet authored 7 years ago
  
  e9635127
Jul 24, 2017
- ae finetuning added · 9d87c506
  Cédric Traizet authored 7 years ago
  
  9d87c506
Jul 20, 2017
- bug correct : output vector of m_net.evalLayer was not taken from the right layer · b315eb89
  Cédric Traizet authored 7 years ago
  
  b315eb89
- switched to linear neuron (test), bug correct in the DoPredictBatch method · 4a75d17a
  Cédric Traizet authored 7 years ago
  
  4a75d17a
- Merge branch 'master' into dev · 7eee49a4
  Cédric Traizet authored 7 years ago
  
  Conflicts: include/AutoencoderModel.txx include/DimensionalityReductionModelFactory.txx include/cbLearningApplicationBaseDR.h
  7eee49a4
- refactoring of the ae model seems to work, denoising ae is not implemented yet · f0b2c8dd
  Cédric Traizet authored 7 years ago
  
  f0b2c8dd
Jul 19, 2017
- switched to linear linear ae · ea00ca90
  Cédric Traizet authored 7 years ago
  
  ea00ca90
- changed the way ae are initialized · fecc26c1
  Cédric Traizet authored 7 years ago
  
  fecc26c1
Jul 18, 2017
- cout error at every epochs during training (for testing purposes) · 082cc389
  Cédric Traizet authored 7 years ago
  
  082cc389
Jul 13, 2017
- vec app problem solved · dc19d1cb
  Cédric Traizet authored 7 years ago
  
  dc19d1cb
- debugging the vector reduction app · 80f9d699
  Cédric Traizet authored 7 years ago
  
  80f9d699
Jul 12, 2017
- neuronType switched to tanh/Linear, some cout removed · e7e3f089
  Cédric Traizet authored 7 years ago
  
  e7e3f089
Jul 06, 2017
- Started refactoring of ae (paused) · b2fc854c
  Cédric Traizet authored 7 years ago
  
  b2fc854c
- some temporary cout added for testing purposes · af9aeea9
  Cédric Traizet authored 7 years ago
  
  af9aeea9
Jul 05, 2017
- PCA now outputs Eigenvectors and eigenvalues in a txt file · 0887295c
  Cédric Traizet authored 7 years ago
  
  0887295c
- Autoencoder now outputs the encoder matrix and the hidden bias if asked · 1b8af99c
  Cédric Traizet authored 7 years ago
  
  1b8af99c
- SOM can now output the trained map in a txt file · 9cb87753
  Cédric Traizet authored 7 years ago
  
  9cb87753
Jul 03, 2017
- BUG : Serialization of SOM models works again · ad08b84f
  Cédric Traizet authored 7 years ago
  
  ad08b84f
Jun 30, 2017

BUG : SOMModel now herits from MachineLearningModel<TInput,...> instead of... · 47fc69c1

Cédric Traizet authored 7 years ago

BUG : SOMModel now herits from MachineLearningModel<TInput,...> instead of MachineLearningModel<itk:variablelengthvector<TInput>,...>

47fc69c1

Jun 29, 2017
- the learning curve file can now be set in the app · ef7227bd
  Cédric Traizet authored 7 years ago
  
  ef7227bd
- learning curves added, field for points ID added in vector dr app output shapefile · 94ab0bf1
  Cédric Traizet authored 7 years ago
  
  94ab0bf1
Jun 27, 2017
- comments added/modified · 67b19129
  Cédric Traizet authored 7 years ago
  
  67b19129
- Training an autoencoder now outputs a learning curve (txt file), work in progress · 2c525d93
  Cédric Traizet authored 7 years ago
  
  2c525d93
- m_Dimension is now an attribute of ML models, it is now possible to change the dim of a pca model · 649836a8
  Cédric Traizet authored 7 years ago
  
  649836a8
Jun 21, 2017
- DRmodel headers removed (MLmodels are used instead) · 717f6f93
  Cédric Traizet authored 7 years ago
  
  717f6f93
- DR is now based on Machine Learning Models, it is now possible to reduce de... · ea53b867
  Cédric Traizet authored 7 years ago
  
  DR is now based on Machine Learning Models, it is now possible to reduce de dimension of the pca model (it has to be changed in PCAModel.txx for now
  ea53b867
Jun 20, 2017
- som is now a machine learning model, ae and pca have been deactiveted · a30fb6f2
  Cédric Traizet authored 7 years ago
  
  a30fb6f2
Jun 19, 2017
- debug · b108c5b3
  Cédric Traizet authored 7 years ago
  
  b108c5b3
Jun 13, 2017
- Sparse ae added, parameters can now be defined for each layers (noise,... · a79f0d47
  Cédric Traizet authored 7 years ago
  
  Sparse ae added, parameters can now be defined for each layers (noise, regularization, rho and beta)
  a79f0d47
Jun 06, 2017
- vector app work in progress · 00fb23c7
  Cédric Traizet authored 7 years ago
  
  00fb23c7
Jun 02, 2017
- New application for vector dimensionality reductio · 916ac460
  Cédric Traizet authored 7 years ago
  
  916ac460
May 30, 2017
- multi layer autoencoders now working (app and reduction) · bc8b732a
  Cédric Traizet authored 7 years ago
  
  bc8b732a
May 29, 2017
- autoencodermodel's attribute "m_net" is now a vector (works for a vector of size 1) · b2b0fc4e
  Cédric Traizet authored 7 years ago
  
  b2b0fc4e
- Merge branch 'master' into drmodel · cf6f75b9
  Cédric Traizet authored 7 years ago
  
  old master is outdated
  cf6f75b9
- some cmments and documentation added · 82e47317
  Cédric Traizet authored 7 years ago
  
  82e47317