trf::Model_FastSample Class Reference
TRF model, revise the sample method to speedup the MCMC. More...
#include <trf-model.h>
Inheritance diagram for trf::Model_FastSample:
Public Member Functions | |
| Model_FastSample (Vocab *pv) | |
| Model_FastSample (Vocab *pv, int maxlen) | |
| LogP | ProposeW0 (VocabID &wi, Seq &seq, int nPos, bool bSample=true) |
| void | ProposeCProbs (VecShell< LogP > &logps, Seq &seq, int nPos) |
| void | MarkovMove (Seq &seq) |
| [sample] Markov Move - perform the gibbs sampling More... | |
| void | SamplePos (Seq &seq, int nPos) |
 Public Member Functions inherited from trf::Model | |
| Model (Vocab *pv) | |
| constructor More... | |
| Model (Vocab *pv, int maxlen) | |
| constructor More... | |
| ~Model () | |
| destructor More... | |
| int | GetMaxLen () const |
| Get max-len. More... | |
| Vocab * | GetVocab () const |
| Get Vocab. More... | |
| int | GetMaxOrder () const |
| Get maximum order. More... | |
| int | GetParamNum () const |
| Get parameter number. More... | |
| void | Reset (Vocab *pv, int maxlen) |
| reset, the maxlen is the length excluding the beg/end symbols. More... | |
| virtual void | SetParam (PValue *pValue) |
| Set the parameters. More... | |
| void | GetParam (PValue *pValue) |
| Get the paremetre vector. More... | |
| void | SetPi (Prob *pPi) |
| Set the pi. More... | |
| template<typename T > | |
| void | SetZeta (T *pzeta) |
| Set updated zeta. More... | |
| template<typename T > | |
| void | GetZeta (T *pzeta) |
| LogP | GetLogProb (Seq &seq, bool bNorm=true) |
| calculate the probability More... | |
| void | LoadFromCorpus (const char *pcorpus, const char *pfeatstyle, int nOrder) |
| load ngram features from corpus More... | |
| void | FeatCount (Seq &seq, double *pCount, double dadd=1.0) |
| Count the feature number in a sequence. More... | |
| void | ReadT (const char *pfilename) |
| Read Model. More... | |
| void | WriteT (const char *pfilename) |
| Write Model. More... | |
| LogP | ClusterSum (Seq &seq, int nPos, int nOrder) |
| Read Binary. More... | |
| virtual double | ExactNormalize (int nLen) |
| [exact] Exact Normalization, return the logz of given length More... | |
| virtual void | ExactNormalize () |
| [exact] Exact Normalization More... | |
| void | GetNodeExp (int nLen, double *pExp) |
| [exact] E_{p_l}[f]: Exactly calculate the expectation over x and h for length nLen More... | |
| void | GetNodeExp (double *pExp, Prob *pLenProb=NULL) |
| [exact] sum_l { n_l/n * E_{p_l}[f] }: Exactly calculate the expectation over x and h More... | |
| void | Sample (Seq &seq) |
| [sample] Perform one train-dimensional mixture sampling More... | |
| void | LocalJump (Seq &seq) |
| [sample] Local Jump - sample a new length More... | |
| LogP | ProposeLength (int nOld, int &nNew, bool bSample) |
| [sample] Propose the length, using the variable m_matLenJump More... | |
| LogP | ProposeC0 (VocabID &ci, Seq &seq, int nPos, bool bSample) |
| [sample] Propose the c_{i} at position i. Then return the propose probability R(c_i|h_i,c_{other}) More... | |
| void | ProposeCProbs (VecShell< LogP > &logps, Seq &seq, int nPos) |
| [sample] Return the propose distribution of c_i at position nPos More... | |
| LogP | GetReducedModelForC (Seq &seq, int nPos) |
| [sample] A unnormalized reduced model to sample class c_i. More... | |
| LogP | GetReducedModelForW (Seq &seq, int nPos) |
| [sample] A unnormalized reduced model to sample word w_i. More... | |
| LogP | GetReducedModel (Seq &seq, int nPos) |
| [sample] A unnormalized reduced depending on nPos. More... | |
| LogP | GetMarginalProbOfC (Seq &seq, int nPos) |
| [sample] given c_i, summate the probabilities of x_i, i.e. P(c_i) More... | |
| void | SampleC (Seq &seq, int nPos) |
| [sample] Sample the c_i at position nPos without x_i. More... | |
| LogP | SampleX (Seq &seq, int nPos, bool bSample=true) |
| [sample] Sample the x_i at position nPos More... | |
| LogP | AISNormalize (int nLen, int nChain, int nInter) |
| perform AIS to calculate the normalization constants, return the logz of given length More... | |
| void | AISNormalize (int nLenMin, int nLenMax, int nChain, int nInter) |
Public Attributes | |
| int | m_nMHtimes |
| Public Attributes inherited from trf::Model | |
| Feat * | m_pFeat |
| hash all the features More... | |
| Vec< PValue > | m_value |
| the value for each features More... | |
| int | m_maxlen |
| the maximum length of model, excluding <s> and </s>. The min-len = 1 More... | |
| Vec< Prob > | m_pi |
| the prior length distribution More... | |
| Vec< LogP > | m_logz |
| the normalization constants log Z_l More... | |
| Vec< LogP > | m_zeta |
| the estimated normalization constants (fix = 0) More... | |
| Vocab * | m_pVocab |
| Mat< Prob > | m_matLenJump |
| [sample] used to propose a new length More... | |
| int | m_maxSampleLen |
| [sample] the maximum sample length, default = m_maxlen + 2 More... | |
| int | m_nLenJumpAccTimes |
| lenght jump the acceptance times More... | |
| int | m_nLenJumpTotalTime |
| total times of length jump More... | |
| int | m_nSampleHAccTimes |
| sample H the acceptance times More... | |
| int | m_nSampleHTotalTimes |
Additional Inherited Members | |
| Protected Attributes inherited from trf::Model | |
| AlgNode | m_AlgNode |
| the forward-backward calculation each node More... | |
Detailed Description
TRF model, revise the sample method to speedup the MCMC.
Definition at line 217 of file trf-model.h.
Constructor & Destructor Documentation
§ Model_FastSample() [1/2]
|
inline |
Definition at line 222 of file trf-model.h.
§ Model_FastSample() [2/2]
|
inline |
Definition at line 225 of file trf-model.h.
Member Function Documentation
§ MarkovMove()
|
inlinevirtual |
[sample] Markov Move - perform the gibbs sampling
Reimplemented from trf::Model.
Definition at line 276 of file trf-model.h.
§ ProposeCProbs()
Definition at line 265 of file trf-model.h.
§ ProposeW0()
|
inline |
Definition at line 228 of file trf-model.h.
§ SamplePos()
|
inline |
Definition at line 281 of file trf-model.h.
Member Data Documentation
§ m_nMHtimes
| int trf::Model_FastSample::m_nMHtimes |
Definition at line 220 of file trf-model.h.
The documentation for this class was generated from the following file:
- C:/Users/zhang/spmi/SPMILM/tools/trf/src/TRF/trf-model.h
