fineTuneDArch,DArch-method

A instance of the class <code><a href="/link/?package=darch&version=0.12.0" darch'="" data-mini-rdoc="darch::"></a>'>DArch</code>.

darch

<code><a href="/link/?package=darch&version=0.12.0" dataset'="" data-mini-rdoc="darch::"></a>'>DataSet</code> containing training and 
optionally validation and test data.

dataSet

<code><a href="/link/?package=darch&version=0.12.0" dataset'="" data-mini-rdoc="darch::"></a>'>DataSet</code> to be used for validation.

dataSetValid

numEpochs

Indicates whether the training is for a classification net. 
When <code>TRUE</code> then statistics for classification will be determind. 
Default is <code>TRUE</code>

isClass

Stop criteria for the error on the train data. Default is 
<code>-Inf</code>

stopErr

Stop criteria for the classification error on the train 
data. Default is <code>101</code>

stopClassErr

Stop criteria for the error on the validation data. 
Default is <code>-Inf</code>.

stopValidErr

Stop criteria for the classification error on the 
validation data. Default is <code>101</code>.

stopValidClassErr

Whether to shuffle train data before each epoch.

shuffleTrainData

Whether to enable debug mode, internal parameter.

debugMode

Additional parameters for the training function

Fine tuning function for the deep architecture

internal

The darch package is built on the basis of the code from G. E.
Hinton and R. R. Salakhutdinov (available under Matlab Code for deep belief
nets). This package is for generating neural networks with many layers (deep
architectures) and train them with the method introduced by the publications
"A fast learning algorithm for deep belief nets" (G. E. Hinton, S. Osindero,
Y. W. Teh (2006) <DOI:10.1162/neco.2006.18.7.1527>) and "Reducing the
dimensionality of data with neural networks" (G. E. Hinton, R. R.
Salakhutdinov (2006) <DOI:10.1126/science.1127647>). This method includes a
pre training with the contrastive divergence method published by G.E Hinton
(2002) <DOI:10.1162/089976602760128018> and a fine tuning with common known
training algorithms like backpropagation or conjugate gradients.
Additionally, supervised fine-tuning can be enhanced with maxout and
dropout, two recently developed techniques to improve fine-tuning for deep
learning.

Martin Drees

Package for Deep Architectures and Restricted Boltzmann Machines

fineTuneDArch,DArch-method function

A instance of the class <code><a href='<a href='DArch'></a>'>DArch</a></code>.

<code><a href='<a href='DataSet'></a>'>DataSet</a></code> containing training and 
optionally validation and test data.

<code><a href='<a href='DataSet'></a>'>DataSet</a></code> to be used for validation.

fineTuneDArch,DArch-method: Fine tuning function for the deep architecture

Description

Usage

Arguments

See Also