I followed tutorial "Training Acoustic Model For CMUSphinx" (http://cmusphinx.sourceforge.net/wiki/tutorialam), I followed all the steps and I run

./scripts_pl/RunAll.pl

Script was completed all tasks.

But I have errors when I run decoder:

./scripts_pl/decode/slave.pl

ERROR: "ngram_model_arpa.c", line 465: File /task/etc/task.lm.DMP not found
ERROR: "ngram_model_dmp.c", line 105: Dump file /task/etc/task.lm.DMP not found
ERROR: "ngram_search.c", line 210: Failed to read language model file: /task/etc/task.lm.DMP

How i should create task.lm.DMP file ? Scripts does not create it automatically?

Or I missed something ?

Best regards.

--- (Edited on 12/3/2010 9:43 am [GMT-0600] by last_deceneu) ---

Hello

File .lm.DMP is a language model in DMP format. You can learn how to create them reading other part of tutorial

http://cmusphinx.sourceforge.net/wiki/tutoriallm

To create a language model you can use test prompts as a source.

Btw, the tutorial you are reading mentions that in this part

http://cmusphinx.sourceforge.net/wiki/tutorialam#data_preparation

Please read it carefully

Please use CMUSphinx forums to ask questions about CMUSPhinx.

--- (Edited on 12/3/2010 19:08 [GMT+0300] by nsh) ---

Thank you for your help.

This models names are so similar and so easy to miss-understand.

--- (Edited on 12/3/2010 2:55 pm [GMT-0600] by last_deceneu) ---

You are welcome

Please suggest corrections to cleanup the tutorial. That would be much appreciated.

--- (Edited on 12/4/2010 01:40 [GMT+0300] by nsh) ---

Hello all!

I'm verry happy to contribute with everythink.

My problem rised because I'm not verry familiar with all terminologyes of ASR.

I tutorial I founded

   etc
     your_db.dic
     your_db.phone
     your_db.lm.DMP
     your_db.fillers
     your_db_train.fileids
     your_db_train.transcription
     your_db_test.fileids
     your_db_test.transcription</code>

then later was explanations:

Language model file should be in ARPA format or in DMP format.

Filler dictionary contains filler phones (not-covered by language model non-linguistic sounds like breath,
hmm or laugh). It can contain just silences.

Here I was in touble to identify Language model and related file.

I suggest for each term specify exactly what is related file ex.  (your_db.lm.DMP)

--- (Edited on 12/6/2010 3:12 am [GMT-0600] by last_deceneu) ---