First perform PCFG training on the generated Nuance grammar, defined
by the {\tt ebl\_\-nuance\_\-grammar} or {\tt nuance\_\-grammar} config file
entry. The training data is taken from the file defined by the
{\tt ebl\_\-grammar\_\-probs} config file entry.
Next, compile the PCFG-trained version of the Nuance grammar, produced
by the first step, into a recognition package with the same name. This
will be done using the Nuance language pack defined by the
{\tt nuance\_\-language\_\-pack} config file entry and the extra parameters
defined by the {\tt nuance\_\-compile\_\-params} config file entry. Typical
values for these parameters are as follows:
\begin{verbatim}
regulus_config(nuance_language_pack, 'English.America').
regulus_config(nuance_compile_params,
['-auto_pron', '-dont_flatten']).
\end{verbatim}
