I would like to use "offline" tools to generate a dictionary and language
model for use with our system.
We would like to change the dictionary / language model dynamically based on
the context as we are using a command & control type system.
We would also like to generate the dictionary & lm dynamically.
I have downloaded quick_lm.pl which is used by the lmtool, but this does not
seem to run on my machine, from the source, I cannot see that quick_lm
generates a dictionary, so I am guessing there are more scripts / tools that
do this.
I also looked at the cmuclmtk, which I guess will generate my language model,
but not my dictionary.
Ideally I need to generate a concise dictionary from my language model or
similar.
Can someone please shed some light on this as the documentation is pretty
sparse and I can't find anything in the various source directories that seems
to imply that this is possible.
Thanks
Glen
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
Thanks, I had read / have been re-reading those tutorials, but my problem is I
want to start from a corpus of words.
The tutorials keep referring me to the LM tool if I want to use a corpus of
words. Also, the quick_lm tool does not seem to be the "complete" tool that
lmtool provides and I think the downloadable version is out of date, or
different to the one the tool actually runs - I can't seem to even get it to
run on my machine, but that may be my problem not the file from here:
Hi,
I would like to use "offline" tools to generate a dictionary and language
model for use with our system.
We would like to change the dictionary / language model dynamically based on
the context as we are using a command & control type system.
We would also like to generate the dictionary & lm dynamically.
I have downloaded quick_lm.pl which is used by the lmtool, but this does not
seem to run on my machine, from the source, I cannot see that quick_lm
generates a dictionary, so I am guessing there are more scripts / tools that
do this.
I also looked at the cmuclmtk, which I guess will generate my language model,
but not my dictionary.
Ideally I need to generate a concise dictionary from my language model or
similar.
Can someone please shed some light on this as the documentation is pretty
sparse and I can't find anything in the various source directories that seems
to imply that this is possible.
Thanks
Glen
Hello
Please check the tutorial
http://cmusphinx.sourceforge.net/wiki/tutorial
In particular
http://cmusphinx.sourceforge.net/wiki/tutoriallm
http://cmusphinx.sourceforge.net/wiki/tutorialdict
Hi,
Thanks, I had read / have been re-reading those tutorials, but my problem is I
want to start from a corpus of words.
The tutorials keep referring me to the LM tool if I want to use a corpus of
words. Also, the quick_lm tool does not seem to be the "complete" tool that
lmtool provides and I think the downloadable version is out of date, or
different to the one the tool actually runs - I can't seem to even get it to
run on my machine, but that may be my problem not the file from here:
http://www.speech.cs.cmu.edu/tools/download/quick_lm.pl
Thanks
Glen
Sorry, maybe I'm living a different world. I doublechecked the link and it
never mentions a quick_lm. You should not use quick_lm at all.
Instead the tutorial mentions a pack of handy tools like phonetisaurus, srilm,
pronounce, mitlm, cmuclmtk. Maybe you'll find it interesting to read.