For a while, I've been using the LMTools website to create my Language Model
and Dictionary files. Recently, however, I've found it necessary to compile
them locally due to their size. My problem is that whenever I run it locally,
the resulting LM (.arpa) file causes my application to crash. If I can take
the same file and run in through LMTools is works fine. Is there something I'm
missing here?
Here are the commands I'm running on my local machine:
Please provide the versions of the software you are using.
Please try to use latest versions. For cmuclmtk please use subversion snapshot
Please when you report about crash try to provide the log of the crash, the list of the arguments you were passing. If you describe the problems in details it's way easier to solve it.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi All -
For a while, I've been using the LMTools website to create my Language Model
and Dictionary files. Recently, however, I've found it necessary to compile
them locally due to their size. My problem is that whenever I run it locally,
the resulting LM (.arpa) file causes my application to crash. If I can take
the same file and run in through LMTools is works fine. Is there something I'm
missing here?
Here are the commands I'm running on my local machine:
cat a.text | ./text2wfreq > a.wfreq
cat a.wfreq | ./wfreq2vocab -top 20000 > a.vocab
cat a.text | ./text2idngram -temp /tmp -vocab a.vocab -write_ascii <a.text>
a.idngram
./idngram2lm -idngram a.idngram -vocab a.vocab -arpa a.arpa -vocab_type 0
-good_turing -disc_ranges 0 0 0 -ascii_input </a.text>
Thanks, Andrew J Durstewitz
http://cmusphinx.sourceforge.net/wiki/tutoriallm#building_statistical_languag
e_model_using_cmuclmtk
Please provide the versions of the software you are using.
Please try to use latest versions. For cmuclmtk please use subversion snapshot
Please when you report about crash try to provide the log of the crash, the list of the arguments you were passing. If you describe the problems in details it's way easier to solve it.