hrest.tex

来自「隐马尔可夫模型源代码」· TEX 代码 · 共 171 行

%/* ----------------------------------------------------------- */
%/*                                                             */
%/*                          ___                                */
%/*                       |_| | |_/   SPEECH                    */
%/*                       | | | | \   RECOGNITION               */
%/*                       =========   SOFTWARE                  */ 
%/*                                                             */
%/*                                                             */
%/* ----------------------------------------------------------- */
%/*         Copyright: Microsoft Corporation                    */
%/*          1995-2000 Redmond, Washington USA                  */
%/*                    http://www.microsoft.com                */
%/*                                                             */
%/*   Use of this software is governed by a License Agreement   */
%/*    ** See the file License for the Conditions of Use  **    */
%/*    **     This banner notice must not be removed      **    */
%/*                                                             */
%/* ----------------------------------------------------------- */
%
% HTKBook - Steve Young    31/10/95
%

\newpage
\mysect{HRest}{HRest}

\mysubsect{Function}{HRest-Function}

\index{hrest@\htool{HRest}|(}
\htool{HRest} performs basic Baum-Welch re-estimation of the parameters of
a single HMM using a set of observation sequences.  \htool{HRest} can be used
for normal isolated word training in which the observation sequences
are realisations of the corresponding vocabulary word.  

Alternatively,
\htool{HRest} can be used to generate \textit{seed} HMMs for phoneme-based
recognition.  In this latter case, the observation sequences will consist
of segments of continuously spoken training material.  \htool{HRest} will
\textit{cut} these out of the training data automatically by simply
giving it a segment label.

In both of the above applications, \htool{HRest} is intended to operate on
HMMs with initial parameter values estimated by \htool{HInit}.

\htool{HRest} supports multiple mixture components, multiple streams,
parameter tying within a single model, full or diagonal
covariance matrices, tied-mixture models and discrete models.  
The outputs of \htool{HRest} are often further 
processed by \htool{HERest}.

Like all re-estimation tools, \htool{HRest} allows a floor to  be set on
each individual variance by defining a variance floor macro for each
data stream (see chapter~\ref{c:Training}).  If any diagonal covariance
component falls below $0.00001$, then the corresponding mixture weight is
set to zero.  A warning is issued if the number of mixtures is greater
than one, otherwise an error occurs.  Applying a variance floor via the
{\tt -v} option or a variance floor macro can be used to prevent this.

\mysubsect{Use}{HRest-Use}

\htool{HRest} is invoked via the command line
\begin{verbatim}
   HRest [options] hmm trainFiles ...
\end{verbatim}
This causes the parameters of the given {\tt hmm} to be
re-estimated repeatedly using the data in {\tt trainFiles} 
until either a maximum iteration limit is reached
or the re-estimation converges. 
The HMM definition can be contained within one or more macro
files loaded via the standard \texttt{-H} option.  Otherwise, the 
definition will be read from a file called \texttt{hmm}.
The list of train files
can be stored in a script file if required.

The detailed operation of \htool{HRest} is controlled by the following
command line options
\begin{optlist}
  \ttitem{-c f} Set the threshold for tied-mixture observation
      pruning to {\tt f}.
      When all mixtures of all models are tied to create a full
      tied-mixture system, the calculation of output probabilities
      is treated as a special case.  Only those 
      mixture component probabilities which fall within {\tt f} of
      the maximum mixture component probability are used in calculating
      the state output probabilities (default 10.0).

  \ttitem{-e f} This sets the convergence factor to the real value {\tt f}.
      The convergence factor is the relative change between successive
      values of $P({O}|\lambda)$ (default value 0.0001).

  \ttitem{-i N} This sets the maximum number of re-estimation cycles
      to {\tt N} (default value 20).

  \ttitem{-l s} The string {\tt s} must be the name of a
      segment label.  When this option is used, \htool{HRest} searches
      through all of the training files and cuts out all segments with
      the given label.  When this option is not used, \htool{HRest} assumes that
      each training file is a single token.

  \ttitem{-m N}  Sets the minimum number of training examples to be {\tt N}.
      If fewer than {\tt N} examples are supplied then an error is
      reported (default value 3).

  \ttitem{-t}  Normally, training sequences are rejected if they
      have fewer frames
      than the number of emitting states in the HMM.  Setting this
      switch disables this reject mechanism\footnote{Using this option
      only makes sense if the HMM has skip transitions}.
      
  \ttitem{-u flags}  By default, \htool{HRest} updates all 
      of the HMM parameters,
      that is, means, variances, mixture weights
      and transition probabilities. This 
      option causes just the parameters indicated by the {\tt flags}
      argument to be updated, this argument is a string containing one
      or more of the letters {\tt m} (mean), {\tt v} (variance), 
      {\tt t} (transition) and {\tt w} (mixture weight).  The presence of a
      letter enables the updating of the corresponding parameter set.

  \ttitem{-v f}  This sets the minimum variance (i.e. diagonal element of
      the covariance matrix) to the real value {\tt f}.  This is ignored
      if an explicit variance floor macro is defined. The default value
      is 0.0.
      
  \ttitem{-w f}  Any mixture weight or discrete observation probability
      which falls below the global
            constant {\tt MINMIX} is treated as being zero.
      When this parameter is  set,  all mixture weights  are floored
      to {\tt f * MINMIX}.

\stdoptB
\stdoptF
\stdoptG
\stdoptH
\stdoptI
\stdoptL
\stdoptM
\stdoptX

\end{optlist}
\stdopts{HRest}

\mysubsect{Tracing}{HRest-Tracing}

\htool{HRest} supports the following trace options where each
trace flag is given using an octal base
\begin{optlist}
   \ttitem{000001} basic progress reporting.
   \ttitem{000002} output information on the training data loaded.
   \ttitem{000004} the observation probabilities.
   \ttitem{000010} the alpha matrices.
   \ttitem{000020} the beta matrices.
   \ttitem{000040} the occupation counters.
   \ttitem{000100} the transition counters.
   \ttitem{000200} the mean counters.
   \ttitem{000400} the variance counters.
   \ttitem{001000} the mixture weight counters.
   \ttitem{002000} the re-estimated transition matrix.
   \ttitem{004000} the re-estimated mixture weights.
   \ttitem{010000} the re-estimated means.
   \ttitem{020000} the re-estimated variances.
\end{optlist}
Trace flags are set using the \texttt{-T} option or the  \texttt{TRACE} 
configuration variable.
\index{hrest@\htool{HRest}|)}


%%% Local Variables: 
%%% mode: latex
%%% TeX-master: "../htkbook"
%%% End: