I can run pocketsphinx_continuous correctly under Ubuntu system but after
cross-compiled over to ARM9 it gave me the following error:
looks like bio.c (line 89) cannot find end_comment from bio_readhdr ... can
anyone give me a quick hint on what might go wrong or the difference while
cross-compile? Thanks very much!
=======
INFO: acmod.c(238): Parsed model-specific feature parameters from
/envoasr/envo_lm/rm1.cd_cont_1000/feat.params
It looks like your 'transition_matrices' file was corrupted for some reason.
It doesn't start with a proper header I think:
s3
version 1.0
chksum0 yes
endhdr
If you trained your model on windows, it might have windows-style newlines.
It's actually very easy to solve your problem, you just need to debug your
application a little and find the reason of the probem. You can easily add
printf in mentioned point in the code, you can easily find the reason of wrong
behavior as well.
The function and source file are provided in message.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Anonymous
-
2010-05-20
Resolved - as you mentioned it's a file corruption issue after trace the code.
Thanks!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I can run pocketsphinx_continuous correctly under Ubuntu system but after
cross-compiled over to ARM9 it gave me the following error:
looks like bio.c (line 89) cannot find end_comment from bio_readhdr ... can
anyone give me a quick hint on what might go wrong or the difference while
cross-compile? Thanks very much!
=======
INFO: acmod.c(238): Parsed model-specific feature parameters from
/envoasr/envo_lm/rm1.cd_cont_1000/feat.params
INFO: feat.c(848): Initializing feature stream to type: '1s_c_d_dd',
ceplen=13, CMN='current', VARNORM='no', AGC='none'
INFO: cmn.c(142): mean= 12.00, mean= 0.0
INFO: mdef.c(520): Reading model definition:
/envoasr/envo_lm/rm1.cd_cont_1000/mdef
INFO: bin_mdef.c(173): Allocating 32703 * 8 bytes (255 KiB) for CD tree
INFO: tmat.c(205): Reading HMM transition probability matrices:
/envoasr/envo_lm/rm1.cd_cont_1000/transition_matrices
FATAL_ERROR: "bio.c", line 89: Missing end_comment
marker
It looks like your 'transition_matrices' file was corrupted for some reason.
It doesn't start with a proper header I think:
If you trained your model on windows, it might have windows-style newlines.
It's actually very easy to solve your problem, you just need to debug your
application a little and find the reason of the probem. You can easily add
printf in mentioned point in the code, you can easily find the reason of wrong
behavior as well.
The function and source file are provided in message.
Resolved - as you mentioned it's a file corruption issue after trace the code.
Thanks!