Learn R Programming

⚠️There's a newer version (0.12.0) of this package.Take me there.

darch (version 0.9.1)

Package for deep architectures and Restricted-Bolzmann-Machines

Description

The darch package is build on the basis of the code from G. E. Hinton and R. R. Salakhutdinov (available under Matlab Code for deep belief nets : last visit: 01.08.2013). 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) and "Reducing the dimensionality of data with neural networks" (G. E. Hinton, R. R. Salakhutdinov). This method includes a pre training with the contrastive divergence method publishing by G.E Hinton (2002) and a fine tuning with common known training algorithms like backpropagation or conjugate gradient.

Copy Link

Version

Monthly Downloads

17

Version

0.9.1

License

GPL-2

Issues

Pull Requests

Stars

Forks

Repository

http://github.com/maddin79/darch

Maintainer

Martin Drees

Last Published

March 16th, 2014

Functions in darch (0.9.1)

Returns the execution output of the layer from the DArch object

Constructor function for RBM object.

Binary sigmoid unit function.

Returns the number of visible units of the RBM

Deep architectures in R

getGenWeightFunction

Returns the function for generating weight matrices.

Returns the cancel value.

Returns the list of statistics for the network

Pre trains a DArch network

getPosPhaseData

Returns the data for the positive phaes.

getLearnRateBiasVisible

Returns the learning rate for the visible biases.

Adds a field to a layer

setErrorFunction<-

Sets the error function of the Net.

crossEntropyError

Cross entropy error function

Abtract class for neural networks.

getLearnRateBiases

Returns the learning rate for the bias weigths of the DArch object.

getExecuteFunction

Returns the function for the execution of the DArch network.

Resets the weights and biases of the RBM object

Returns if the weights are saved as ff objects

Function for generating ff files of the MNIST Database

Calculates the linear neuron output no transfer function

getHiddenBiases

Returns the biases of the hidden units.

Minimize a differentiable multivariate function.

Returns the execution output list of the DArch object

Class for Restricted-Bolzmann-Machine

getErrorFunction

Returns the error function of the Net.

Fine tuning function for the deep architecture.

getFineTuneFunction

Returns the fine tune function for the DArch object.

minimizeAutoencoder

Conjugate gradient for a autoencoder network

generateWeights

Generates a weight matrix.

Generates the rbm's for the pre-training.

getCancelMessage

Returns the cancel message.

Returns the output of the RBM

Adds a layer to the DArch object

setPosPhaseData<-

Sets the positive phase data for the training

setExecuteFunction<-

Sets the execution function for the network

Sets the layers for the network

setFinalMomentum<-

Sets the final momentum of the Net.

sigmoidUnitDerivative

Sigmoid unit function with unit derivatives.

Function for updating the weights and biases of an RBM

Returns the field of a layer from the DArch object.

getLearnRateBiasHidden

Returns the learning rate for the hidden biases.

getFinalMomentum

Returns the final momentum of the Net.

getHiddenBiasesInc

Returns the update value for the biases of the hidden units.

getLearnRateWeights

Returns the learn rate of the weights.

Returns a list of RBMs of the DArch object.

Sets the momentum of the Net.

Returns the momentum of the Net.

getVisibleBiasesInc

Returns the update value for the biases of the visible units.

softmaxUnitDerivative

Softmax unit function with unit derivatives.

getLayerWeights

Returns the weights of a layer from the DArch object.

Sets the log level for the Net.

Mean quared error function

Returns the a list of layers from the DArch object.

setVisibleBiasesInc<-

Sets the update value for the biases of the visible units

Saves a RBM network

Adds an execution output for a DArch object

setNumVisible<-

Sets the number of visible units

Returns the update value for the weights.

Returns the number of hidden units of the RBM

linearUnitDerivative

Linear unit function with unit derivatives.

setLearnRateBiasVisible<-

Sets the learnig rates of the biases for the visible units

sigmUnitFuncSwitch

Calculates the neuron output with the sigmoid function

getVisibleBiases

Returns the biases of the visible units.

Loads a RBM network

Saves a DArch network

Constructor function for DArch objects.

Execute the darch

Sets a layer with the given index for the network

setHiddenBiasesInc<-

Sets the update value for the biases of the hidden units

setHiddenBiases<-

Sets the biases of the hidden units for the RBM object

minimizeClassifier

Conjugate gradient for a classification network

setCancelMessage<-

Sets the cancel message.

setHiddenUnitFunction<-

Sets the unit function of the hidden units

setLearnRateBiasHidden<-

Sets the learnig rates of the biases for the hidden units

loadRBMFFWeights

Loads weigths and biases for a RBM network from a ffData file.

Sets the list of RBMs

Sets the number of hidden units

removeLayerField

Removes a layer from the DArch object

Loads a DArch network

setMomentumSwitch<-

Sets the momentum switch of the Net.

setLayerFunction<-

Sets the function for a layer with the given index

Returns the weigth cost for the training

Resilient-Backpropgation training for deep architectures.

setVisibleUnitFunction<-

Sets the unit function of the visible units

setFineTuneFunction<-

Sets the fine tuning function for the network

Sets the weights of the RBM object

getVisibleUnitStates

Returns a list with the states of the visible units.

makeStartEndPoints

Makes start- and end-points for the batches.

getLayerFunction

Returns the neuron function of a layer from the DArch object.

backpropagation

Backpropagation learning function

Linear unit function.

Sigmoid unit function.

setVisibleBiases<-

Sets the biases of the visible units for the RBM object

Sets the batch size of the Net.

resetExecOutput

Resets the output list of the DArch object

Sets the update values for the weights

setLearnRateWeights<-

Sets the learning rate for the weights.

Softmax unit function.

Adds a list of statistics to the network

Set whether the learning shall be canceled.

setLayerWeights<-

Sets the weights of a layer with the given index

setHiddenUnitStates<-

Sets the states of the hidden units

setWeightCost<-

Sets the weight costs for the training

Quadratic error function

setVisibleUnitStates<-

Sets the states of the visible units

Resets the weights and biases of the DArch object

Sets if the weights are saved as ff objects

Returns the weigths of the RBM.

saveRBMFFWeights

Saves weights and biases of a RBM network into a ffData file.

setGenWeightFunction<-

Sets the function for generating weight matrices.

setUpdateFunction<-

Sets the update function of the RBM object

Returns the batch size of the Net.

Sets the output of the RBM object

setLearnRateBiases<-

Sets the learning rate for the biases

Trains a RBM with contrastive divergence

Returns a layer from the DArch object.

getHiddenUnitStates

Returns a list with the states of the hidden units.

getMomentumSwitch

Returns the momentum switch of the Net.

Calculates the neuron output with the sigmoid function

setLayerField<-

Sets a field in a layer.

Returns the logger of the Net.

Sets the logger of the Net.