Hi,

I created Tamil Language models using sphinxTrain. when i decode with sphinx3 i'm getting following error

MODULE: DECODE Decoding using models previously trained
        Decoding 24 segments starting at 0 (part 1 of 1)
        0%
This step had 25 ERROR messages and 27 WARNING messages.  Please check the log file for details.
        Aligning results to find error rate
        SENTENCE ERROR: 100.0% (24/24)   WORD ERROR RATE: 100.0% (53/53)

Errors in log file are

INFO: lm_3g_dmp.c(471): Bad magic number: 1735287116(676e614c), not an LM dumpfile??
ERROR: "lm_3g_dmp.c", line 1274: Error in reading the header of the DUMP file.

And many instances of following error

WARNING: "vithist.c", line 784: No word exit in frame 534, using exits from frame 113
ERROR: "vithist.c", line 811: No word exit in frame 534, using exits from frame 113

I have uploaded the decoding and training log files in the following location.

http://www.mediafire.com/?jv5b4muwhzk1231

http://www.mediafire.com/?j3s2xij5sbkcxik

Please help me.

Thanks!

Hanitha

--- (Edited on 12/17/2010 2:08 pm [GMT-0600] by ghanitha) ---