I found that the sphinx3_lm_convert program is implemented only for converting
trigram models to binary DMP format. If I have 5gram models is there a way to
convert the language model to binary DMP format? Please let me know.
it gives me the following error when i tried to do INFO: lm.c(612): LM
read('sorted', lw= 1.00, wip= 0.10, uw= 1.00) INFO: lm.c(614): Reading LM file
sorted (LM name "default") INFO: lm_3g.c(832): Reading LM file sorted WARNING:
"lm_3g.c", line 250: Unknown ngram (4) WARNING: "lm_3g.c", line 843: Couldnt'
read the ngram count INFO: lm.c(663): LM is not in TXT format FATAL_ERROR:
"main_lm_convert.c", line 184: Fail to read inputfn sorted in inputfmt TXT
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear Sir,
sorry for posting in the wrong forum.
I found that the sphinx3_lm_convert program is implemented only for converting
trigram models to binary DMP format. If I have 5gram models is there a way to
convert the language model to binary DMP format? Please let me know.
it gives me the following error when i tried to do INFO: lm.c(612): LM
read('sorted', lw= 1.00, wip= 0.10, uw= 1.00) INFO: lm.c(614): Reading LM file
sorted (LM name "default") INFO: lm_3g.c(832): Reading LM file sorted WARNING:
"lm_3g.c", line 250: Unknown ngram (4) WARNING: "lm_3g.c", line 843: Couldnt'
read the ngram count INFO: lm.c(663): LM is not in TXT format FATAL_ERROR:
"main_lm_convert.c", line 184: Fail to read inputfn sorted in inputfmt TXT
No, dmp format doesn't support 5-grams.