MODULE: DECODE Decoding using models previously trained Aligning results to find error rate word_align.pl failed with error code 65280 at /usr/local/lib/sphinxtrain/scripts/decode/slave.pl line 173.

[rezaee]

Hi
I wanted to train my own acoustic model. after successful passing MODULE 90 I got this error message:

MODULE: DECODE Decoding using models previously trained
        Aligning results to find error rate
word_align.pl failed with error code 65280 at /usr/local/lib/sphinxtrain/scripts/decode/slave.pl line 173.

How can I fix it?

Also I have another question, my wav files format are something like this:

1/1.wav
1/2.wav
1/3.wav
2/1.wav
2/2.wav
2/3.wav
3/1.wav
3/2.wav
3/3.wav
. 
.
.

Is it a true format?

[rezaee]

Some files were corrupted and also I think I should did change "mymodel.lm.dmp" section in configuration files to "mymodel.bin" as I had .bin file in my database instead of .lm.dmp and it worked

[Ali]

I have same error ?
Could you plz tell me how can I find the corrcupted file ?

also my langugae model file name is .lm only
should I need to replace

$DEC_CFG_LANGUAGEMODEL = "$CFG_BASE_DIR/etc/${CFG_DB_NAME}.lm.DMP";
with

$DEC_CFG_LANGUAGEMODEL = "$CFG_BASE_DIR/etc/${CFG_DB_NAME}.lm";
on my sphinx_train.cfg

[Nickolay V. Shmyrev]

Read decode log in logdir

[Ali]

got the empty log
only date is written on it

[Nickolay V. Shmyrev]

You can not run pocketsphinx because you didn't configure the linker properly.

[Ali]

I am using sphinxbase and sphinxtrain only for training an acoustic model ?
Should i need to run ./configure , make and make intstall command on sphinxbase , sphinxtrain and pocketsphinx folder If yes then in which order

Or
should I need to run ./configure , make and make intstall command on sphinxbase , sphinxtrain only

[Ridwan Bhugaloo]

how to configure the linker properly?

[Ali]

Thanks every one I was not compiling pocketsphinx
I am succusscfully train an acoustic model using an4 database that is provided

Now I need to train it on my own database

Special thank Nickolay V. Shmyrev , for replying my post

[Nickolay V. Shmyrev]

I am succusscfully train an acoustic model using an4 database that is provided

Congratulations, let us know how it goes.