#include <kb.h>
Public Attributes | |
| kbcore_t * | kbcore |
| fe_t * | fe |
| mfcc_t ** | mfcc |
| float32 *** | feat |
| ascr_t * | ascr |
| beam_t * | beam |
| fast_gmm_t * | fastgmm |
| pl_t * | pl |
| adapt_am_t * | adapt_am |
| stat_t * | stat |
| FILE * | matchfp |
| FILE * | matchsegfp |
| FILE * | hmmdumpfp |
| int32 | op_mode |
| char * | uttid |
| char * | uttfile |
| void * | srch |
| adapt_am_t* kb_t::adapt_am |
Structure that wraps up adaptation variables. such as regression matrices in MLLR Structure that wraps up parameters for adaptation such as MLLR.
| ascr_t* kb_t::ascr |
Structures of storing parameters for different techniques. Senone and composite senone scores for one frame.
| beam_t* kb_t::beam |
Structure that wraps up parameters related to beam pruning.
| fast_gmm_t* kb_t::fastgmm |
Structure that wraps up parameters for fast GMM computation.
| fe_t* kb_t::fe |
Feature generation related variables Front-end processing object
| float32*** kb_t::feat |
Dynamic feature frames
| FILE* kb_t::hmmdumpfp |
File handle for dumping hmms for debugging
| kbcore_t* kb_t::kbcore |
Core models, defined as acoustic and language models, dictionary (pronounciation models_, front-ends, filler-penalties approximate acoustic models such as sub vector quantization map and Gaussian selector Core model structures
| FILE* kb_t::matchfp |
FILE handle that handles output. File handle for the match file
Referenced by main().
| FILE* kb_t::matchsegfp |
File handle for the match segmentation file
Referenced by main().
| mfcc_t** kb_t::mfcc |
Input feature frames (may be NULL)
| int32 kb_t::op_mode |
| pl_t* kb_t::pl |
Structure that wraps up parameters for phoneme look-ahead.
| void* kb_t::srch |
The search structure
| char* kb_t::uttfile |
Utterance Filename. The other thing that should move somewhere else.
| char* kb_t::uttid |
A mode for specifying operation Utterance ID. The one thing that should move to somewhere like srch
1.8.11