1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
|
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.77 [en] (X11; U; SunOS 5.7 sun4u) [Netscape]">
</head>
<body text="#000000" bgcolor="#FFFFFF" link="#0000EE" vlink="#551A8B" alink="#FF0000">
<center><b><font size=+4>Torch Reference Manual</font></b>
<br><img SRC="torche.jpeg" ALT="[Torch]" NOSAVE height=32 width=32>
<p> <a href="http://www.torch.ch">http://www.torch.ch</a>
<br><a href="mailto:collober@idiap.ch">collober@idiap.ch</a>
<br><a href="mailto:collober@iro.umontreal.ca">collober@iro.umontreal.ca</a></center>
<br>
</body>
</html>
<H1>Table of Contents</H1>
<H2>General stuff</H2>
<UL>
<LI><A HREF="Listfunctions.html">List functions</A>
<LI><A HREF="Memoryfunctions.html">Memory functions</A>
<LI><A HREF="SeveralfunctionsforInputsOuputsondisk..html">Several functions for Inputs/Ouputs on disk.</A>
<LI><A HREF="Textoutputsfunctions..html">Text outputs functions.</A>
</UL>
<H2>Namespaces</H2>
<UL>
<LI><A HREF="Givensmatrixoperationsroutines..html"> Givens matrix operations routines.</A>
<LI><A HREF="Householdertransformationroutines..html"> Householder transformation routines.</A>
<LI><A HREF="RoutinesfordeterminingHessenbergfactorisations..html"> Routines for determining Hessenberg factorisations.</A>
<LI><A HREF="Routinesforsymmetriceigenvalueproblems..html"> Routines for symmetric eigenvalue problems.</A>
<LI><A HREF="Collectionofmatrixfactorisationoperationfunctions..html">Collection of matrix factorisation operation functions.</A>
<LI><A HREF="Collectionofmatrixoperationfunctions..html">Collection of matrix operation functions.</A>
<LI><A HREF="Collectionofpermutationsoperationfunctions..html">Collection of permutations operation functions.</A>
<LI><A HREF="Randomfunctions..html">Random functions.</A>
<LI><A HREF="Somesimplefunctionsforlogoperations..html">Some simple functions for log operations.</A>
<LI><A HREF="Somesimplefunctionsforstringoperations..html">Some simple functions for string operations.</A>
</UL>
<H2>Classes</H2>
<UL>
<LI><A HREF="Bagging.html">Bagging</A> <I>This class represents a <TT>Trainer</TT> that implements the well-known Bagging algorithm (Breiman, 1996).</I>
<LI><A HREF="BayesClassifier.html">BayesClassifier</A> <I>A multi class bayes classifier -- maximizes the likelihood of each class separately using a trainer for distribution.</I>
<LI><A HREF="BayesClassifierMachine.html">BayesClassifierMachine</A> <I>BayesClassifierMachine is the machine used by the <TT>BayesClassifier</TT> trainer to perform a Bayes Classification using different distributions.</I>
<LI><A HREF="Boosting.html">Boosting</A> <I>Boosting implementation.</I>
<LI><A HREF="BoostingMeasurer.html">BoostingMeasurer</A> <I>Compute the classification error (in %) for <TT>BoostingMachine</TT> of the <TT>inputs</TT> with respect to the <TT>targets</TT> of <TT>data</TT>.</I>
<LI><A HREF="ClassFormat.html">ClassFormat</A> <I>Used to define a class code.</I>
<LI><A HREF="ClassLLCriterion.html">ClassLLCriterion</A> <I>This criterion can be used to train *in classification* a <TT>GradientMachine</TT> object using the <TT>GMTrainer</TT> trainer.</I>
<LI><A HREF="ClassMeasurer.html">ClassMeasurer</A> <I>Compute the classification error (in %) of the <TT>inputs</TT> with respect to the <TT>targets</TT> of <TT>data</TT>.</I>
<LI><A HREF="CmdLine.html">CmdLine</A> <I>This class provides a useful interface for the user, to easily read some arguments/options from the command-line.</I>
<LI><A HREF="ConnectedMachine.html">ConnectedMachine</A> <I>Easy connections between several <TT>GradientMachine</TT>.</I>
<LI><A HREF="Criterion.html">Criterion</A> <I><TT>Criterion</TT> class for <TT>GMTrainer</TT>.</I>
<LI><A HREF="DataSet.html">DataSet</A> <I>Provides an interface to manipulate all kind of data.</I>
<LI><A HREF="DiagonalGMM.html">DiagonalGMM</A> <I>This class can be used to model Diagonal Gaussian Mixture Models.</I>
<LI><A HREF="Dictionary.html">Dictionary</A> <I>This class contains the dictionary of accepted words for a speech recognition experiment, such as the one used by SpeechHMM.</I>
<LI><A HREF="DistrMachine.html">DistrMachine</A> <I>This class can be used to implement a conditional distribution P(y|x;theta).</I>
<LI><A HREF="Distribution.html">Distribution</A> <I>This class is designed to handle generative distribution models such as Gaussian Mixture Models and Hidden Markov Models.</I>
<LI><A HREF="DotKernel.html">DotKernel</A> <I>DotProduct</I>
<LI><A HREF="EMTrainer.html">EMTrainer</A> <I>This class is used to train any distribution using the EM algorithm.</I>
<LI><A HREF="EditDistance.html">EditDistance</A> <I>This class can be used to compute the "edit distance" between two sequences.</I>
<LI><A HREF="EditDistanceMeasurer.html">EditDistanceMeasurer</A> <I>This class can be used to measure and print an <TT>EditDistance</TT> object.</I>
<LI><A HREF="EuclideanDataSet.html">EuclideanDataSet</A> <I><TT>DataSet</TT> with dot-products.</I>
<LI><A HREF="Exp.html">Exp</A> <I>Exponentiel layer for <TT>GradientMachine</TT>.</I>
<LI><A HREF="FileDataSet.html">FileDataSet</A> <I>Create a <TT>DataSet</TT> from a disk file.</I>
<LI><A HREF="FileSparseDataSet.html">FileSparseDataSet</A> <I>Create a <TT>DataSet</TT> from a <EM>sparse</EM> disk file.</I>
<LI><A HREF="FixedMachineDistribution.html">FixedMachineDistribution</A> <I>This class uses one of the outputs of a given pre-trained machine as an estimate of a probability (used in the method <TT>frameLogProbability</TT>.</I>
<LI><A HREF="GMTrainer.html">GMTrainer</A> <I>Trainer for GradientMachine.</I>
<LI><A HREF="GaussianKernel.html">GaussianKernel</A> <I>Gaussian <IMG BORDER=0 SRC=g000007.gif></I>
<LI><A HREF="GradientMachine.html">GradientMachine</A> <I>Gradient machine: machine which can be trained with a gradient descent.</I>
<LI><A HREF="Grammar.html">Grammar</A> <I>This class contains the grammar of accepted sentences for a speech recognition experiment such as the one using SpeechHMM A grammar is a transition table where each node is a word.</I>
<LI><A HREF="HMM.html">HMM</A> <I>This class implements a Hidden Markov Model distribution.</I>
<LI><A HREF="HtkFileDataSet.html">HtkFileDataSet</A> <I>Creates a <TT>StdDataSet</TT> from a disk file in HTK format.</I>
<LI><A HREF="HtkSeqDataSet.html">HtkSeqDataSet</A> <I>This class is used to read Htk seqdatasets</I>
<LI><A HREF="IOHtk.html">IOHtk</A> <I>This class is used to read HTK objects (used for various datasets)</I>
<LI><A HREF="IOTorch.html">IOTorch</A> <I>Load and save file in the Torch format.</I>
<LI><A HREF="InputsSelect.html">InputsSelect</A> <I>Machine which select a block of adjacent inputs, and put them in the outputs.</I>
<LI><A HREF="KNN.html">KNN</A> <I>This machine implements the K-nearest-neighbors (KNN) algorithm.</I>
<LI><A HREF="Kernel.html">Kernel</A> <I>Kernel class.</I>
<LI><A HREF="Kmeans.html">Kmeans</A> <I>This class can be used to do a "kmeans" on a given set of data.</I>
<LI><A HREF="Linear.html">Linear</A> <I>Linear layer for <TT>GradientMachine</TT>.</I>
<LI><A HREF="LogRBF.html">LogRBF</A> <I>LogRBF layer for <TT>GradientMachine</TT>.</I>
<LI><A HREF="LogSigmoid.html">LogSigmoid</A> <I>Log-sigmoid layer for <TT>GradientMachine</TT>.</I>
<LI><A HREF="LogSoftmax.html">LogSoftmax</A> <I>LogSoftmax layer for <TT>GradientMachine</TT>.</I>
<LI><A HREF="MLP.html">MLP</A> <I>This class is a simple interface to the <TT>ConnectedMachine</TT> class that ca be used to build the well-known Multi Layer Perceptron type of neural networks.</I>
<LI><A HREF="Machine.html">Machine</A> <I><TT>Object</TT> which can compute some outputs, given some inputs.</I>
<LI><A HREF="Mat.html">Mat</A> <I>Matrix object.</I>
<LI><A HREF="MatSeqDataSet.html">MatSeqDataSet</A> <I>This class enable to read sequences (hence inherits from <TT>SeqDataSet</TT>) from the classical Torch data format.</I>
<LI><A HREF="Measurer.html">Measurer</A> <I>Used to measure what you want during training/testing.</I>
<LI><A HREF="Mixer.html">Mixer</A> <I>Mixer useful for experts mixtures.</I>
<LI><A HREF="MseCriterion.html">MseCriterion</A> <I>Mean Squared Error criterion.</I>
<LI><A HREF="MseMeasurer.html">MseMeasurer</A> <I>Mean Squared Error measurer.</I>
<LI><A HREF="MultiClassFormat.html">MultiClassFormat</A> <I>Define the multi class code.</I>
<LI><A HREF="MultiCriterion.html">MultiCriterion</A> <I>MultiCriterion can be used to handle multiple criterions.</I>
<LI><A HREF="Multinomial.html">Multinomial</A> <I>This class can be used to model Multinomial Distributions.</I>
<LI><A HREF="NPTrainer.html">NPTrainer</A> <I>Trainer for Non Parametric Machines.</I>
<LI><A HREF="NllCriterion.html">NllCriterion</A> <I>This criterion can be used to train <TT>Distribution</TT> object using the <TT>GMTrainer</TT> trainer.</I>
<LI><A HREF="NllMeasurer.html">NllMeasurer</A> <I>This class measures the negative log likelihood.</I>
<LI><A HREF="Object.html">Object</A> <I>Provides a useful interface for managing options.</I>
<LI><A HREF="OneHotClassFormat.html">OneHotClassFormat</A> <I>Define the one hot class code.</I>
<LI><A HREF="Optimizer.html">Optimizer</A> <I>Optimizer for the <TT>GMTrainer</TT> class.</I>
<LI><A HREF="OutputMeasurer.html">OutputMeasurer</A> <I>This class can be used to save the outputs of a <TT>Trainer</TT> in a file in such a format that it can be read again in <TT>Torch</TT> using the <TT>FileDataSet</TT> class.</I>
<LI><A HREF="ParzenDistribution.html">ParzenDistribution</A> <I>This class can be used to model a Parzen density estimator with a Gaussian kernel:</I>
<LI><A HREF="ParzenMachine.html">ParzenMachine</A> <I>This machine implements the Parzen Window estimator.</I>
<LI><A HREF="Perm.html">Perm</A> <I>Permutation object.</I>
<LI><A HREF="PhonemeSeqDataSet.html">PhonemeSeqDataSet</A> <I>This class is designed to create a dataset based on another dataset using parts only related to a given phoneme</I>
<LI><A HREF="PolynomialKernel.html">PolynomialKernel</A> <I>Polynomial <IMG BORDER=0 SRC=g000006.gif>.</I>
<LI><A HREF="QCCache.html">QCCache</A> <I>"Cache" used by the Quadratic Constrained Trainer (<TT>QCTrainer</TT>).</I>
<LI><A HREF="QCMachine.html">QCMachine</A> <I>"Quadratic Constrained Machine".</I>
<LI><A HREF="QCTrainer.html">QCTrainer</A> <I>Train a <TT>QCMachine</TT>.</I>
<LI><A HREF="RBF.html">RBF</A> <I>This class is a simple interface to the <TT>ConnectedMachine</TT> class that ca be used to build the well-known Radial Basis Function type of neural networks.</I>
<LI><A HREF="SVM.html">SVM</A> <I>Support Vector Machine.</I>
<LI><A HREF="SVMCache.html">SVMCache</A> <I><TT>QCCache</TT> implementation for SVMs.</I>
<LI><A HREF="SVMCacheClassification.html">SVMCacheClassification</A> <I>Cache for SVM classification.</I>
<LI><A HREF="SVMCacheRegression.html">SVMCacheRegression</A> <I>Cache for SVM regression.</I>
<LI><A HREF="SVMClassification.html">SVMClassification</A> <I>SVM in classification.</I>
<LI><A HREF="SVMRegression.html">SVMRegression</A> <I>SVM in regression.</I>
<LI><A HREF="SaturationMeasurer.html">SaturationMeasurer</A> <I>Measure the saturation of a <TT>GradientMachine</TT>.</I>
<LI><A HREF="SeqDataSet.html">SeqDataSet</A> <I>This class defines the standard framework of Sequence Data processing</I>
<LI><A HREF="Sigmoid.html">Sigmoid</A> <I>Sigmoid layer for <TT>GradientMachine</TT>.</I>
<LI><A HREF="SigmoidKernel.html">SigmoidKernel</A> <I>Sigmoid <IMG BORDER=0 SRC=g000008.gif></I>
<LI><A HREF="Softmax.html">Softmax</A> <I>Softmax layer for <TT>GradientMachine</TT>.</I>
<LI><A HREF="SparseDataSet.html">SparseDataSet</A> <I>Sparse Data Set.</I>
<LI><A HREF="SparseLinear.html">SparseLinear</A> <I>Sparse Linear layer for <TT>GradientMachine</TT>.</I>
<LI><A HREF="SpeechHMM.html">SpeechHMM</A> <I>This class implements a special case of Hidden Markov Models that can be used to do connected word speech recognition for small vocabulary, using embedded training.</I>
<LI><A HREF="StdDataSet.html">StdDataSet</A> <I>Standard Data Set.</I>
<LI><A HREF="StochasticGradient.html">StochasticGradient</A> <I>Stochastic Gradient Optimizer for GMTrainer</I>
<LI><A HREF="SumMachine.html">SumMachine</A> <I>This machine simply adds up its input vectors.</I>
<LI><A HREF="TableLookupDistribution.html">TableLookupDistribution</A> <I>This class outputs one of the observations as the logProbability.</I>
<LI><A HREF="Tanh.html">Tanh</A> <I>Tanh layer for <TT>GradientMachine</TT>.</I>
<LI><A HREF="TimeMeasurer.html">TimeMeasurer</A> <I>Measure the time (in seconds) between two <TT>measureIter()</TT> calls.</I>
<LI><A HREF="Trainer.html">Trainer</A> <I>Trainer.</I>
<LI><A HREF="TwoClassFormat.html">TwoClassFormat</A> <I>Define the two class code.</I>
<LI><A HREF="Vec.html">Vec</A> <I>Vector object.</I>
<LI><A HREF="ViterbiTrainer.html">ViterbiTrainer</A> <I>This class is used to train any distribution using the Viterbi algorithm.</I>
<LI><A HREF="WeightedMseCriterion.html">WeightedMseCriterion</A> <I>Similar to <TT>MseCriterion</TT>, but you can put a weight on each example.</I>
<LI><A HREF="WeightedSumMachine.html">WeightedSumMachine</A> <I>Weighted-sum machine.</I>
<LI><A HREF="WordSegMeasurer.html">WordSegMeasurer</A> <I>This class can be used to save the word segmentation of a <TT>SpeechHMM</TT> in a file.</I>
</UL>
<H2>Functions</H2>
<UL>
<LI><A HREF="MSTDVNormalize.html">MSTDVNormalize</A> <I>Compute means and variances for normalizing a matrix.</I>
<LI><A HREF="MSTDVSparseNormalize.html">MSTDVSparseNormalize</A> <I>Compute means and variances for normalizing a <EM>sparse</EM> matrix.</I>
<LI><A HREF="deleteExtractedMeasurers.html">deleteExtractedMeasurers</A> <I>Free memory allocations did by <TT>extractMeasurers()</TT>.</I>
<LI><A HREF="extractMeasurers.html">extractMeasurers</A> <I>Make a table of measurers from a <TT>List</TT>.</I>
<LI><A HREF="getRuntime.html">getRuntime</A> <I>Return the time in CLOCKS_PER_SEC</I>
<LI><A HREF="sparseVectorLength.html">sparseVectorLength</A> <I>Return the size of a sparse vector <TT>line</TT>.</I>
</UL>
<H2>Macros</H2>
<UL>
<LI><A HREF="max.html">max</A> <I>The max function</I>
<LI><A HREF="min.html">min</A> <I>The min function</I>
</UL>
<H2>Enums, Unions, Structs</H2>
<UL>
<LI><A HREF="HTKhdr.html">HTKhdr</A> <I>HTK File Header </I>
<LI><A HREF="List.html">List</A> <I>List structure used in all the library.</I>
<LI><A HREF="SeqExample.html">SeqExample</A> <I>This structure keeps a sequence example</I>
<LI><A HREF="sreal.html">sreal</A> <I>Sparse definition.</I>
</UL>
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.77 [en] (X11; U; SunOS 5.7 sun4u) [Netscape]">
</head>
<body text="#000000" bgcolor="#FFFFFF" link="#0000EE" vlink="#551A8B" alink="#FF0000">
<div align=right><img SRC="torche.jpeg" ALT="[Torch]" NOSAVE height=20 width=20>Torch.
The Ultimate Machine Learning Library.</div>
</body>
</html>
<HR>
<BR>
This page was generated with the help of <A HREF="http://docpp.sourceforge.net">DOC++</A>.
</BODY>
</HTML>
|
