Sphinxtrain does not continue without error after generating mfc files

[wunna]

I try to create a new acoustic model for my language. I used the latest versions of sphinxtrain, sphinxbase and pocketsphinx: sphinxtrain-5prealpha, sphinxbase-5prealpha and pocketsphinx-5prealpha. When I train my model, the sphinxtrain stop after phase 6 Checking that all the words in the transcript are in the dictionary. My dictionary has a lot of duplicated phones because of the nature of our language, Myanmar. I shared my logdir folder.
[https://www.dropbox.com/s/6qcl9ndl6wpqcuy/logdir.rar?dl=0]In that folder, only 000.comp_feat folder contains. Please advise and help me.

[wunna]

This is my logdir folder.

[Nickolay V. Shmyrev]

Phones must be unique, you should also fix other warnings reported.

There is nothing in Myanmar different from other languages, you still can properly select a phoneset according to the wikipedia page

https://en.wikipedia.org/wiki/Burmese_language

You need to share your etc folder if you need help on the database setup, logdir folder is not enough since all your mistakes are in etc folder.

[wunna]

Some words and syllables have two different phones and I faced a lot of warnings in this case.
This is etc folder.

[Nickolay V. Shmyrev]

Please open the wikipedia page linked above and read about phonemes of your language. Please do not use syllables as phones.

[wunna]

There is no phonetic dictionary in Myanmar. Therefore, I try to created one syllable to one phone mapping. To create phonetic dictionary, MLC (Myanmar Language Commession) dictionary has over 200000 words and I cannot write those words to phonetic dictionary manually. I also used Phonetisaurus tools to create g2p mapping but the results are wrong. So I created phonetic dictionary for Myanmar syllables and there are over 2000 syllables in Myanmar. Please guide me for creating phonetic dictionary.

[Nickolay V. Shmyrev]

Your dictionary should look like this:

၁   T I'
၂   H N I'
၃   TH O U N:
၄   L E I:
၅   NG A:
၆   CH A U'
၇   KH U N
၈   SH I'
၉   K O U:
၀   TH O U N N J A.
က   K A.
ကာ  K A
ကား K A:
ကိ  K I.
ကီ  K I
ကီး K I:
ကု  K U.
ကူ  K U
ကူး K U:
ေက  K E I
ကဲ  K E:
ကဲ့ K E.
ေကာ K O:
ေကာ့    K O.
ေကာ္    K O
ကံ  K A N
ကံ့ K A N.
က့ံ K A N.
ကို K O U
ကိုး    K O U:
ကက္ K E'
ကုက္    K O U'
ေကာက္   K A U'
ကိုက္   K A I'
ကင္ K I N
ကင္း    K I N:
ေကာင္   K A U N

Phones must be separated by spaces. Syllables are not phones.

[wunna]

Now I created new phonetic dictionary and phones files. I setup my model in Ubuntu 12.04 at VMware. But when I train my model, he sphinxtrain stop after phase 6 Checking that all the words in the transcript are in the dictionary. At this stage, I faced many warnings.Can those warnings stop training process?
I posted my etc folder and logdir folder.

[Nickolay V. Shmyrev]

All the words from the train transcription must be in a dictionary

[wunna]

This is logdir folder.