Hi,
I am trying to train the model using sphinx. I have create my own set of
dictionary, language model. The input file is in the .wav format. Each file
contains only one word per file. I am trying to build model based on the input
format.
I was able to train the model. But when I typed
perl RunAll.pl
I am getting the following errors. These errors are produced in the log file.
Can anyone please help me with this process.
I paste only the part of the error here. Can anyone please tell me how to deal
with this ?
INFO: ........\src\programs\bw\main.c(421): Silence Tag SIL
INFO: ........\src\libs\libio\corpus.c(1343): Will process all remaining
utts starting at 0
INFO: ........\src\programs\bw\main.c(620): Reestimation: Baum-Welch
WARNING: "........\src\programs\bw\main.c", line 774: utt recordedVoice_01
too short
utt> 0 recordedVoice_02 1WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_02 too short
utt> 0 recordedVoice_03 1WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_03 too short
utt> 0 recordedVoice_04 4WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_04 too short
utt> 0 recordedVoice_05 6WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_05 too short
utt> 0 recordedVoice_06 2WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_06 too short
utt> 0 recordedVoice_07 3WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_07 too short
utt> 0 recordedVoice_08 2WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_08 too short
utt> 0 recordedVoice_09 3WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_09 too short
utt> 0 recordedVoice_010 3WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_010 too short
utt> 0 recordedVoice_011 4WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_011 too short
utt> 0 recordedVoice_012 9 0INFO:
........\src\libs\libcommon\cvt2triphone.c(199): no multiphones defined, no
conversion done
28 6 ERROR: "........\src\programs\bw\backward.c", line 431: final state
not reached
ERROR: "........\src\programs\bw\baum_welch.c", line 331:
user/recordedVoice_012 ignored
Thank you very much....
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Most likely reason of the failure is incorrect format for the recorded data.
Audio files must be in MSWAV format, 16 khz 16 bit mono single channel.
According to the tutorial, if you want to ask for help about training, try to
provide the training folder. You can pack it into archive, upload to a public
file sharing resource and give here a link to it.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
I am trying to train the model using sphinx. I have create my own set of
dictionary, language model. The input file is in the .wav format. Each file
contains only one word per file. I am trying to build model based on the input
format.
I was able to train the model. But when I typed
perl RunAll.pl
I am getting the following errors. These errors are produced in the log file.
Can anyone please help me with this process.
I paste only the part of the error here. Can anyone please tell me how to deal
with this ?
INFO: ........\src\programs\bw\main.c(421): Silence Tag SIL
INFO: ........\src\libs\libio\corpus.c(1343): Will process all remaining
utts starting at 0
INFO: ........\src\programs\bw\main.c(620): Reestimation: Baum-Welch
WARNING: "........\src\programs\bw\main.c", line 774: utt recordedVoice_01
too short
utt> 0 recordedVoice_02 1WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_02 too short
utt> 0 recordedVoice_03 1WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_03 too short
utt> 0 recordedVoice_04 4WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_04 too short
utt> 0 recordedVoice_05 6WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_05 too short
utt> 0 recordedVoice_06 2WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_06 too short
utt> 0 recordedVoice_07 3WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_07 too short
utt> 0 recordedVoice_08 2WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_08 too short
utt> 0 recordedVoice_09 3WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_09 too short
utt> 0 recordedVoice_010 3WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_010 too short
utt> 0 recordedVoice_011 4WARNING: "........\src\programs\bw\main.c", line
774: utt recordedVoice_011 too short
utt> 0 recordedVoice_012 9 0INFO:
........\src\libs\libcommon\cvt2triphone.c(199): no multiphones defined, no
conversion done
28 6 ERROR: "........\src\programs\bw\backward.c", line 431: final state
not reached
ERROR: "........\src\programs\bw\baum_welch.c", line 331:
user/recordedVoice_012 ignored
Thank you very much....
Read the troubleshooting section of the training tutorial
http://cmusphinx.sourceforge.net/wiki/tutorialam#troubleshooting
Most likely reason of the failure is incorrect format for the recorded data.
Audio files must be in MSWAV format, 16 khz 16 bit mono single channel.
According to the tutorial, if you want to ask for help about training, try to
provide the training folder. You can pack it into archive, upload to a public
file sharing resource and give here a link to it.