new model pocketspinx5 prealpha

[Marcello]

I creating a model in Italian with about 300 words, I do not have to create speeches, only recognize key words, I use the Ubuntu operating system and I created it in the /usr/local/share/pocketsphinx/model/ ara my new acoustic model by creating the etc folder and the waw folder where i have a speaker1 folder inside and a 16Hz mono wav file when I prepare the database i put it in the model folder: /usr/local/share/pocketsphinx/model/ara and type the command: sphinxtrain -t an4 setup , as from tutorials(https://cmusphinx.github.io/wiki/tutorialam/) I should have the new structure:

├─ etc
├─ feat
├─ logdir
├─ model_parameters
├─ model_architecture
├─ result
└─ wav

but this does not happen, two files are created in etc: feat.params and sphinx_train.cfg , my start structure is:
├─ etc
│ ├─ ara.dic
│ ├─ ara.phone
│ ├─ ara.filler
│ ├─ ara_train.fileids
│ ├─ ara_train.transcription
│ ├─ ara_test.fileids
│ └─ ara_test.transcription
│ └─ ara.jsgf
└─ wav
├─ speaker_1
└─ file_1.wav

 i setup the file sphinx_train.cfg, with add comment a line: $CFG_LANGUAGEMODEL, and remove comment at line : DEC_CFG_GRAMMAR, and setup $CFG_INITIAL_DENSITIES and $CFG_FINAL_NUM_DENSITIES =8,  now if sphinxtrain -t an4 run, in log file: sphinx_fe.c(967): Processing all remaining utterances at position 0

INFO: sphinx_fe.c(787): Converting /usr/local/share/pocketsphinx/model/ara/wav/speaker1/file_1.wav to /usr/local/share/pocketsphinx/model/ara/feat/speaker1/file_1.mfc
ERROR: "sphinx_fe.c", line 118: Failed to open /usr/local/share/pocketsphinx/model/ara/wav/speaker1/file_1.wav: No such file or directory, but the file wav exist in this directory!!

my model is here https://github.com/McNamara10/ara

can you help me?

[Nickolay V. Shmyrev]

The folder with wav files must be called "wav", not "waw" with final 'w' like in your github. When computer tells you the file is missing check yourself, computers do not lie or try to fool you.

[Marcello]

yes I have see now, the error is stupid..pardon :), ow converting, but have another error in phone list, so after command run if you process fine, i will a new structure: ├─ etc
├─ feat
├─ logdir
├─ model_parameters
├─ model_architecture
├─ result
└─ wav

or i need other operations?

[Nickolay V. Shmyrev]

It is hard to understand your question.

[Marcello]

sorry, but i not understand because i launc this command: sphinxtrain -t ara setup , not created folders: ├─model_parameters
├─ model_architecture
├─ result
as an example--> https://cmusphinx.github.io/wiki/tutorialam/

so i have convert my file.wav in mfc, (with your help..the folder was waw..i update at wav)
now sphinxtrain -t ara run is completed,

MODULE: 000 Computing feature from audio files  (2017-10-24 10:24)

Extracting features from segments starting at (part 1 of 1)

sphinx_fe Log File
completed
Extracting features from segments starting at (part 1 of 1)

sphinx_fe Log File
completed
Feature extraction is done

MODULE: 00 verify training files    (2017-10-24 10:24)

Phase 1: Checking to see if the dict and filler dict agrees with the phonelist file.
WARNING: The phonelist (/usr/local/share/pocketsphinx/model/ara/etc/ara.phone) does not define the phone SIL (required!)
Found 4 words using 6 phones
WARNING: This phone (SIL) occurs in the dictionary (/usr/local/share/pocketsphinx/model/ara/etc/ara.dic), but not in the phonelist (/usr/local/share/pocketsphinx/model/ara/etc/ara.phone)
passed
Phase 2: Checking to make sure there are not duplicate entries in the dictionary
passed
Phase 3: Check general format for the fileids file; utterance length (must be positive); files exist
passed
Phase 4: Checking number of lines in the transcript file should match lines in fileids file
passed
Phase 5: Determine amount of training data, see if n_tied_states seems reasonable.
Estimated Total Hours Training: 0.000425
This is a small amount of data, no comment at this time
WARNING
Phase 6: Checking that all the words in the transcript are in the dictionary
Words in dictionary: 1
Words in filler dictionary: 3
passed
Phase 7: Checking that all the phones in the transcript are in the phonelist, and all phones in the phonelist appear at least once
passed

i have see ara.html from folder my model (ara), it's correct, my new structure is this: https://github.com/McNamara10/model-ara-new/tree/master/ara

if this is correct, i launch this command:
perl scripts_pl/000.comp_feat/slave_feat.pl
perl scripts_pl/00.verify/verify_all.pl
perl scripts_pl/10.vector_quantize/slave.VQ.pl
perl scripts_pl/20.ci_hmm/slave_convg.pl
perl scripts_pl/30.cd_hmm_untied/slave_convg.pl
perl scripts_pl/40.buildtrees/slave.treebuilder.pl
perl scripts_pl/45.prunetree/slave-state-tying.pl
perl scripts_pl/50.cd_hmm_tied/slave_convg.pl
perl scripts_pl/90.deleted_interpolation/deleted_interpolation.pl
as an example --> https://cmusphinx.github.io/wiki/tutorialam/

and this script created this file?
mdef
feat.params
mixture_weights
means
noisedict
transition_matrices
variances

[Marcello]

now i understand, i see the log..i have correct the error, and with command run, has created the folder, but i have error..,
INFO: feat.c(715): Initializing feature stream to type: '1s_c_d_dd', ceplen=13, CMN='batch', VARNORM='no', AGC='none'
ERROR: "acmod.c", line 79: Folder '/usr/local/share/pocketsphinx/model/ara/model_parameters/ara.cd_cont_10' does not contain acoustic model definition 'mdef'
FATAL: "batch.c", line 822: PocketSphinx decoder init failed