tips:speech:simon

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Next revision		 Previous revision
tips:speech:simon [2017/03/04 18:09] – created scipiotips:speech:simon [2017/03/14 20:59] (current) – [compile] scipio
Line 1: Line 1:
 ====== simon ====== ====== simon ======
 +
 +speech model basically consists of two parts:
 +  * Language model: Describes all existing words and what sentences are grammatically correct. It will be created from your active scenario
 +  * Acoustic model: Describes how words sound. It will be either built solely through your voice recordings (training) or with the help of a base model
 +
 +
 +SPHINX backend" uses:
 +  * PocketSphinx recognizer
 +  * SphinxTrain model generation backend
 +
 +models:
 +  * US English HUB4 WSJ SPHINX model. This model is part of pocketsphinx-extra where it is called \"hub4_wsj_sc_3s_8k.cd_semi_5000\".
 +===== install sphinx =====
 +
 +prereq
 +<code bash>
 +sudo apt install bison gradle openjdk-8-jdk
 +</code>
 +
 +build from sources http://cmusphinx.sourceforge.net/wiki/download, not use ubuntu packages
 +
 +
 +
  
 ===== install ===== ===== install =====
 +
 +
 +
  
 <code bash> <code bash>
 sudo apt install simon sudo apt install simon
  
-sudo apt install libsphinxbase3 python-sphinxbase sphinxbase-utils pocketsphinx+sudo apt install libsphinxbase3 python-sphinxbase sphinxbase-utils pocketsphinx sphinxtrain
 </code> </code>
 +
 +edit 
 +<file ini /usr/lib/sphinxtrain/etc/sphinx_train.cfg> 
 +$CFG_CD_TRAIN = 'no';
 +</file>
 +
 +===== compile =====
 +
 +https://phabricator.kde.org/source/simon/
 +
 +<code bash>
 +sudo apt-get build-dep simon
 +
 +git clone git://anongit.kde.org/simon
 +cd simon
 +export QT_VERSION=qt4
 +rm -fR build
 +./build_ubuntu.sh
 +</code>
 +
 +
 +==== model ====
 +
 +=== adapting ===
 +
 +http://cmusphinx.sourceforge.net/wiki/tutorialadapt
 +
 +<code bash>
 +sudo apt install pocketsphinx-en-us/xenial
 +
 +</code>
 +
 +<code bash>
 +mkdir /tmp/model
 +cd /tmp/model
 +cp -a /usr/share/pocketsphinx/model/en-us/en-us .
 +cp /usr/share/pocketsphinx/model/en-us/cmudict-en-us.dict .
 +cp /usr/share/pocketsphinx/model/en-us/en-us.lm.bin .
 +</code>
 +
 +
 +
 +
 +=== training ===
 +
 +
 +training from http://cmusphinx.sourceforge.net/wiki/tutorialamdown
 +
 +* download [[model|http://www.speech.cs.cmu.edu/databases/an4/an4_sphere.tar.gz]]
 +* enter and sphinxtrain -t an4 setup
 +
 +
 +
 +
 +==== old ====
 +
 +For small amount of data for training without context-dependent models use continuous model instead of semi-continuous one and this error will be solved.
 +
 +This warning is caused by the fact you disabled context-dependent training and used semi-continuous model. You can ignore this error until you have more data and use context-dependent training.
 +
 +
 +training
 +
 +<code bash>
 +mkdir /tmp/test
 +cd /tmp/test
 +
 +sphinxtrain -t test setup
 +</code>
 +
 +now we have
 +<code bash>
 +.
 +└── etc
 +│   ├── feat.params
 +│   ├── sphinx_train.cfg
 +│   ├── test.dic
 +│   └── test.filler
 +
 +</code>
 +
 +adjust sphinx_train.cfg TODO
 +
 +
 +
  • tips/speech/simon.1488647377.txt.gz
  • Last modified: 2017/03/04 18:09
  • by scipio