Edit Message

Re: error on recognition by Julius

>After having hmm12 diretory I decided to test with julius to recognize

>the against some test files.

I've never tried this... I have tried against the hmm9 directory from Step 8 (monophone acoustic models), and hmm15 directory hmms from Step 10, and had no problems (though the hmm9 models were not that accurate...)

>### Recognition: 1st pass (LR beam with word-pair grammar)

>Error: 428th frame: no nodes left in beam! model mismatch or wrong

>input?

have you tried modifying the beam width parameters in Julius:

       -b beamwidth
              Beam width (number of HMM nodes) on the first pass.  This  value
              defines  search  width  on the 1st pass, and has great effect on
              the total processing time.  Smaller  width  will  speed  up  the
              decoding,  but  too  small  value  will  result in a substantial
              increase of recognition errors due to  search  failure.   Larger
              value  will make the search stable and will lead to failure-free
              search, but processing time and memory usage will grow  in  pro-
              portion to the width.
 
              Default value is acoustic model dependent:
                400 (monophone)
                800 (triphone,PTM)
               1000 (triphone,PTM, setup=v2.1)

Or why not just finish the acoustic model training and use tied-state triphones?