add to en-us.lm.dmp language model?

[john absalon]

Is there some tool I can use to manually edit the .dmp language model? Or is there some tool I can use to add to the language model? More specifically I would like to combine my custom language model with the general en-us.lm.dmp language model

[bic-user]

.dmp is a binary format of language model. To convert it to editable arpa format use

sphinx_lm_convert -i model.dmp -o model.lm

from sphinxbase.

[Daniel Wolf]

I just downloaded the latest US English generic language model from the SourceForge site. I tried converting the .dmp file using the sphinx_lm_convert.exe tool from the latest pocketsphinx release (5prealpha from 2015-08-05). However, the tool didn't accept the binary format:

sphinx_lm_convert.exe -i cmusphinx-5.0-en-us.lm.dmp -o cmusphinx-5.0-en-us.lm

Current configuration:
[NAME]          [DEFLT] [VALUE]
-case
-debug                  0
-help           no      no
-i                      cmusphinx-5.0-en-us.lm.dmp
-ifmt
-lm_trie        no      no
-logbase        1.0001  1.000100e+00
-mmap           no      no
-o                      cmusphinx-5.0-en-us.lm
-ofmt

INFO: ngram_model_trie.c(456): Trying to read LM in trie binary format
INFO: ngram_model_trie.c(467): Header doesn't match
INFO: ngram_model_trie.c(189): Trying to read LM in arpa format
INFO: ngram_model_trie.c(70): No \data\ mark in LM file
INFO: ngram_model_trie.c(548): Trying to read LM in DMP format
INFO: ngram_model_trie.c(630): ngrams 1=19794, 2=1377200, 3=3178194
INFO: lm_trie.c(399): Training quantizer
INFO: lm_trie.c(407): Building LM trie
ERROR: "ngram_model.c", line 194: language model file type not supported
ERROR: "sphinx_lm_convert.c", line 197: Failed to write language model in format (null) to cmusphinx-5.0-en-us.lm

Does pocketsphinx use different file formats?

[Pedropablo]

Hi! I just came with the same problem and manage to solve it ! Try this:

sphinx_lm_convert -i yourfile.lm.dmp -o yourOutputFile.lm -ofmt arpa

You can check the parameters and arguments if you type :

sphinx_lm_convert -help