Skip to content

Third party:SpeechRecognition:Models:fr

Bertrand Benoit edited this page Dec 7, 2019 · 2 revisions

title: Third-party:SpeechRecognition:Models:fr permalink: /Third-party:SpeechRecognition:Models:fr/

TOC

LIUM models

Follow these instructions, to setup the French models provided by LIUM.

Language model

  • download language model
  • untar trigram_LM.tar, and gunzip trigram_LM.DMP.gz
  • (only if you have compiled sphinxbase) update temporarily the PATH and LD_LIBRARY_PATH to use your version of sphinx_lm_convert (adapt the path if needed)

export LD_LIBRARY_PATH=[HEMERA_TP_PATH](/Appendix#HEMERA_TP_PATH]]/_fromSource/sphinxbase/src/libsphinxbase/.libs:$LD_LIBRARY_PATH export PATH=[HEMERA_TP_PATH](/Appendix#HEMERA_TP_PATH]]/_fromSource/sphinxbase/src/sphinx_lmtools/.libs:$PATH 

  • convert language model file in UTF-8 (using sphinx_lm_convert which supports utf8 encoding, which is NOT the case of sphinx3_lm_convert)

sphinx_lm_convert -i trigram_LM.DMP -ifmt DMP -ienc iso8859-1 -o trigram_LM.DMP.utf8 -oenc utf8 -ofmt DMP

  • move language model file

 mkdir -p [HEMERA_TP_PATH](/Appendix#HEMERA_TP_PATH]]/speechRecognition/data/models/language/lium/3g/  mv trigram_LM.DMP.utf8 [HEMERA_TP_PATH](/Appendix#HEMERA_TP_PATH]]/speechRecognition/data/models/language/lium/3g/

  • Check/update the following configuration elements:

 hemera.core.speechRecognition.sphinx3.languageModel.*

Lexical model

iconv -f iso88591 -t utf8 fillers_dict > fillers_dict.utf8 iconv -f iso88591 -t utf8 words_dict > words_dict.utf8

  • move lexical model files

mkdir -p [HEMERA_TP_PATH](/Appendix#HEMERA_TP_PATH]]/speechRecognition/data/models/lexical/lium/ mv fillers_dict.utf8 words_dict.utf8 [HEMERA_TP_PATH](/Appendix#HEMERA_TP_PATH]]/speechRecognition/data/models/lexical/lium/

  • Check/update the following configuration elements:

hemera.core.speechRecognition.sphinx3.lexicalModel.*

Acoustic model

  • download acoustic model
  • uncompress in [HEMERA_TP_PATH](/Appendix#HEMERA_TP_PATH]]/speechRecognition/data/models/acoustic
  • rename lium_acoustic_models as lium
  • Check/update the following configuration elements:

hemera.core.speechRecognition.sphinx3.acousticModel.*

Category:HemeraBook/en Category:admin/en

You can’t perform that action at this time.