The HTK Book Steve Young Gunnar Evermann Dan Kershaw ...
The HTK Book Steve Young Gunnar Evermann Dan Kershaw ... The HTK Book Steve Young Gunnar Evermann Dan Kershaw ...
17.17 HRest 27717.17 HRest17.17.1 FunctionHRest performs basic Baum-Welch re-estimation of the parameters of a single HMM using a setof observation sequences. HRest can be used for normal isolated word training in which theobservation sequences are realisations of the corresponding vocabulary word.Alternatively, HRest can be used to generate seed HMMs for phoneme-based recognition. Inthis latter case, the observation sequences will consist of segments of continuously spoken trainingmaterial. HRest will cut these out of the training data automatically by simply giving it a segmentlabel.In both of the above applications, HRest is intended to operate on HMMs with initial parametervalues estimated by HInit.HRest supports multiple mixture components, multiple streams, parameter tying within asingle model, full or diagonal covariance matrices, tied-mixture models and discrete models. Theoutputs of HRest are often further processed by HERest.Like all re-estimation tools, HRest allows a floor to be set on each individual variance bydefining a variance floor macro for each data stream (see chapter 8). If any diagonal covariancecomponent falls below 0.00001, then the corresponding mixture weight is set to zero. A warning isissued if the number of mixtures is greater than one, otherwise an error occurs. Applying a variancefloor via the -v option or a variance floor macro can be used to prevent this.17.17.2 UseHRest is invoked via the command lineHRest [options] hmm trainFiles ...This causes the parameters of the given hmm to be re-estimated repeatedly using the data intrainFiles until either a maximum iteration limit is reached or the re-estimation converges. TheHMM definition can be contained within one or more macro files loaded via the standard -H option.Otherwise, the definition will be read from a file called hmm. The list of train files can be stored ina script file if required.The detailed operation of HRest is controlled by the following command line options-c f Set the threshold for tied-mixture observation pruning to f. When all mixtures of all modelsare tied to create a full tied-mixture system, the calculation of output probabilities is treated asa special case. Only those mixture component probabilities which fall within f of the maximummixture component probability are used in calculating the state output probabilities (default10.0).-e f This sets the convergence factor to the real value f. The convergence factor is the relativechange between successive values of P (O|λ) (default value 0.0001).-i N This sets the maximum number of re-estimation cycles to N (default value 20).-l s The string s must be the name of a segment label. When this option is used, HRest searchesthrough all of the training files and cuts out all segments with the given label. When thisoption is not used, HRest assumes that each training file is a single token.-m N Sets the minimum number of training examples to be N. If fewer than N examples are suppliedthen an error is reported (default value 3).-t Normally, training sequences are rejected if they have fewer frames than the number of emittingstates in the HMM. Setting this switch disables this reject mechanism 13 .-u flags By default, HRest updates all of the HMM parameters, that is, means, variances,mixture weights and transition probabilities. This option causes just the parameters indicatedby the flags argument to be updated, this argument is a string containing one or more ofthe letters m (mean), v (variance), t (transition) and w (mixture weight). The presence of aletter enables the updating of the corresponding parameter set.13 Using this option only makes sense if the HMM has skip transitions
17.17 HRest 278-v f This sets the minimum variance (i.e. diagonal element of the covariance matrix) to the realvalue f. This is ignored if an explicit variance floor macro is defined. The default value is 0.0.-w f Any mixture weight or discrete observation probability which falls below the global constantMINMIX is treated as being zero. When this parameter is set, all mixture weights are flooredto f * MINMIX.-B Output HMM definition files in binary format.-F fmt Set the source data format to fmt.-G fmt Set the label file format to fmt.-H mmf Load HMM macro model file mmf. This option may be repeated to load multiple MMFs.-I mlf This loads the master label file mlf. This option may be repeated to load several MLFs.-L dir Search directory dir for label files (default is to search current directory).-M dir Store output HMM macro model files in the directory dir. If this option is not given, thenew HMM definition will overwrite the existing one.-X ext Set label file extension to ext (default is lab).HRest also supports the standard options -A, -C, -D, -S, -T, and -V as described in section 4.4.17.17.3 TracingHRest supports the following trace options where each trace flag is given using an octal base000001 basic progress reporting.000002 output information on the training data loaded.000004 the observation probabilities.000010 the alpha matrices.000020 the beta matrices.000040 the occupation counters.000100 the transition counters.000200 the mean counters.000400 the variance counters.001000 the mixture weight counters.002000 the re-estimated transition matrix.004000 the re-estimated mixture weights.010000 the re-estimated means.020000 the re-estimated variances.Trace flags are set using the -T option or the TRACE configuration variable.
- Page 235 and 236: 16.9 Language modelling tracing 226
- Page 237 and 238: 16.11 Compile-time configuration pa
- Page 239 and 240: Part IVReference Section230
- Page 241 and 242: 17.1 Cluster 23217.1 Cluster17.1.1
- Page 243 and 244: 17.1 Cluster 23417.1.3 TracingClust
- Page 245 and 246: 17.2 HBuild 23617.2.3 TracingHBuild
- Page 247 and 248: 17.3 HCompV 238-l s The string s mu
- Page 249 and 250: 17.4 HCopy 240-t n Set the line wid
- Page 251 and 252: 17.5 HDMan 24217.5 HDMan17.5.1 Func
- Page 253 and 254: 17.5 HDMan 24417.5.3 TracingHDMan s
- Page 255 and 256: 17.6 HEAdapt 246-m f Set the minimu
- Page 257 and 258: 17.7 HERest 24817.7 HERest17.7.1 Fu
- Page 259 and 260: 17.7 HERest 250-w f Any mixture wei
- Page 261 and 262: 17.8 HHEd 252For example,stateComp
- Page 263 and 264: 17.8 HHEd 254FA varscaleComputes an
- Page 265 and 266: 17.8 HHEd 256where V s is the dimen
- Page 267 and 268: 17.8 HHEd 258mixture(m) all mixture
- Page 269 and 270: 17.9 HInit 26017.9 HInit17.9.1 Func
- Page 271 and 272: 17.10 HLEd 26217.10 HLEd17.10.1 Fun
- Page 273 and 274: 17.10 HLEd 26417.10.3 TracingHLEd s
- Page 275 and 276: 17.12 HLMCopy 26617.12 HLMCopy17.12
- Page 277 and 278: 17.13 HLRescore 268-p f Set the wor
- Page 279 and 280: 17.14 HLStats 27017.14.3 UseHLStats
- Page 281 and 282: 17.15 HParse 272Note that C style c
- Page 283 and 284: 17.15 HParse 274HParse will then re
- Page 285: 17.16 HQuant 276where vqFile is the
- Page 289 and 290: 17.18 HResults 280WORD: %Corr=63.91
- Page 291 and 292: 17.18 HResults 28217.18.3 TracingHR
- Page 293 and 294: 17.20 HSLab 28417.20 HSLab17.20.1 F
- Page 295 and 296: 17.20 HSLab 286Load Load a speech d
- Page 297 and 298: 17.21 HSmooth 28817.21 HSmooth17.21
- Page 299 and 300: 17.22 HVite 29017.22 HVite17.22.1 F
- Page 301 and 302: 17.22 HVite 292-v f Enable word end
- Page 303 and 304: 17.23 LAdapt 29417.23.3 TracingLAda
- Page 305 and 306: 17.25 LFoF 29617.25 LFoF17.25.1 Fun
- Page 307 and 308: 17.26 LGCopy 298-r s Set the root n
- Page 309 and 310: 17.28 LGPrep 30017.28 LGPrep17.28.1
- Page 311 and 312: 17.28 LGPrep 302-r s Set the root n
- Page 313 and 314: 17.30 LMerge 30417.30 LMerge17.30.1
- Page 315 and 316: 17.32 LNorm 30617.32 LNorm17.32.1 F
- Page 317 and 318: 17.33 LPlex 30817.33.3 TracingLPlex
- Page 319 and 320: Chapter 18Configuration VariablesTh
- Page 321 and 322: 18.1 Configuration Variables used i
- Page 323 and 324: 18.2 Configuration Variables used i
- Page 325 and 326: 19.1 Generic Errors 316HCopy 1000-1
- Page 327 and 328: 19.2 Summary of Errors by Tool and
- Page 329 and 330: 19.2 Summary of Errors by Tool and
- Page 331 and 332: 19.2 Summary of Errors by Tool and
- Page 333 and 334: 19.2 Summary of Errors by Tool and
- Page 335 and 336: 19.2 Summary of Errors by Tool and
17.17 HRest 278-v f This sets the minimum variance (i.e. diagonal element of the covariance matrix) to the realvalue f. This is ignored if an explicit variance floor macro is defined. <strong>The</strong> default value is 0.0.-w f Any mixture weight or discrete observation probability which falls below the global constantMINMIX is treated as being zero. When this parameter is set, all mixture weights are flooredto f * MINMIX.-B Output HMM definition files in binary format.-F fmt Set the source data format to fmt.-G fmt Set the label file format to fmt.-H mmf Load HMM macro model file mmf. This option may be repeated to load multiple MMFs.-I mlf This loads the master label file mlf. This option may be repeated to load several MLFs.-L dir Search directory dir for label files (default is to search current directory).-M dir Store output HMM macro model files in the directory dir. If this option is not given, thenew HMM definition will overwrite the existing one.-X ext Set label file extension to ext (default is lab).HRest also supports the standard options -A, -C, -D, -S, -T, and -V as described in section 4.4.17.17.3 TracingHRest supports the following trace options where each trace flag is given using an octal base000001 basic progress reporting.000002 output information on the training data loaded.000004 the observation probabilities.000010 the alpha matrices.000020 the beta matrices.000040 the occupation counters.000100 the transition counters.000200 the mean counters.000400 the variance counters.001000 the mixture weight counters.002000 the re-estimated transition matrix.004000 the re-estimated mixture weights.010000 the re-estimated means.020000 the re-estimated variances.Trace flags are set using the -T option or the TRACE configuration variable.