sphinx_lm_convert problem

[Samuel Kitono]

I want to convert an arpa language model to a DMP file but I encounter this
problem...

any suggestion?

58-b0-35-60-f2-f3:bin samuelkitono$ sphinx_lm_convert -i 3kwordsnew.lm -o 3kwordnew.dmp 
INFO: cmd_ln.c(512): Parsing command line:
sphinx_lm_convert \
    -i 3kwordsnew.lm \
    -o 3kwordnew.dmp

Current configuration:
[NAME]      [DEFLT] [VALUE]
-case           
-debug          0
-help       no  no
-i          /Users/samuelkitono/Documents/VocalKit/VocalKitTest/model/lm/5kwords/3kwordsnew.lm
-ienc           
-ifmt           
-logbase    1.0001  1.000100e+00
-mmap       no  no
-o          /Users/samuelkitono/Documents/VocalKit/VocalKitTest/model/lm/5kwords/3kwordnew.dmp
-oenc       utf8    utf8
-ofmt

INFO: ngram_model_arpa.c(476): ngrams 1=2853, 2=1627563, 3=2788528
INFO: ngram_model_arpa.c(135): Reading unigrams
INFO: ngram_model_arpa.c(515):     2853 = #unigrams created
INFO: ngram_model_arpa.c(194): Reading bigrams
........................INFO: ngram_model_arpa.c(531):  1627563 = #bigrams created
INFO: ngram_model_arpa.c(532):    44704 = #prob2 entries
INFO: ngram_model_arpa.c(539):    13400 = #bo_wt2 entries
INFO: ngram_model_arpa.c(291): Reading trigrams
..........................................INFO: ngram_model_arpa.c(552):  2788528 = #trigrams created
INFO: ngram_model_arpa.c(553):    48909 = #prob3 entries
INFO: ngram_model_dmp.c(492): Building DMP model...
INFO: ngram_model_dmp.c(522):     2853 = #unigrams created
Assertion failed: (tgptr - model->lm3g.trigrams < newbase->n_counts[2]), function ngram_model_dmp_build, file ngram_model_dmp.c, line 594.
Abort trap

[Nickolay V. Shmyrev]

This is a bug in sphinxbase which was fixed in trunk. See

https://sourceforge.net/projects/cmusphinx/forums/forum/5471/topic/3646293