semi-continuous models as trained by today sphinxtrain with default setup
continuous models as trained by today sphinxtrain with default setup
However, if you are talking about default hub4_wsj_8k model, you will not be
able to use it because it uses specific feature extraction methods which are
not supported by s2 (transform dct). Also it uses sendump file in specific
format not supported s2. You need to get the version of this model with
mixture weights.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello,
Is there any utility the convert the acoustic models used by pocketsphinx into
sphinx2 format.
Regards
Pankaj
First of all, s2 supports several model types
https://sourceforge.net/projects/cmusphinx/files/sphinxtrain/1.0/SphinxTrain-
1.0.tar.bz2/download
However, if you are talking about default hub4_wsj_8k model, you will not be
able to use it because it uses specific feature extraction methods which are
not supported by s2 (transform dct). Also it uses sendump file in specific
format not supported s2. You need to get the version of this model with
mixture weights.