That is a link to the database that I am using to create my own acoustic
models. My problem is I can't get past Module 20. Something about a file
length mismatch. When I try removing the terminating newline, I just get a
'truncated' error. I am running a WIndows 7 machine with a 64-bit i3
processor. All the text files are in ANSI format with UNIX EOLs and a single
newline at the end of each file. I am using an ARPA language model. Please
tell me what I'm doing wrong. I would appreciate a hasty, but correct
response. I understand I only have a few samples, but this is only going to be
used as a proof of concept. I've searched google and your forums for a
solution, but could not find one. I've tried using the an4 database and I
managed to get past module 20, though I hit another snag, but right now, I'd
just like to get past Module 20. Thanks in advance.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
You still need to cleanup your newlines. There is some issue with them. Please
update the link to the files, it's broken. Then I can tell you what exactly is
wrong.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Anonymous
-
2011-04-04
sorry, it was still uploading, but it's there now. I appreciate the quick
response and hope we can resolve this soon. with regards to your advice, can
you explain what you mean by cleaning up my newlines?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The thing is that you need to point transcription for each file in fileids.
Utterance id is only for check, it's not used to map fileids to transcription.
Fileids file and transcription file should have identical number of lines. You
need to repeat current transcription same number of times as many speakers are
in your database.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Anonymous
-
2011-04-04
thanks so very much for your help. because of you, I have a shot at passing my
thesis. thank you!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
http://dl.dropbox.com/u/11065606/filitext.zip
That is a link to the database that I am using to create my own acoustic
models. My problem is I can't get past Module 20. Something about a file
length mismatch. When I try removing the terminating newline, I just get a
'truncated' error. I am running a WIndows 7 machine with a 64-bit i3
processor. All the text files are in ANSI format with UNIX EOLs and a single
newline at the end of each file. I am using an ARPA language model. Please
tell me what I'm doing wrong. I would appreciate a hasty, but correct
response. I understand I only have a few samples, but this is only going to be
used as a proof of concept. I've searched google and your forums for a
solution, but could not find one. I've tried using the an4 database and I
managed to get past module 20, though I hit another snag, but right now, I'd
just like to get past Module 20. Thanks in advance.
Hello
You still need to cleanup your newlines. There is some issue with them. Please
update the link to the files, it's broken. Then I can tell you what exactly is
wrong.
sorry, it was still uploading, but it's there now. I appreciate the quick
response and hope we can resolve this soon. with regards to your advice, can
you explain what you mean by cleaning up my newlines?
Hello
The thing is that you need to point transcription for each file in fileids.
Utterance id is only for check, it's not used to map fileids to transcription.
Fileids file and transcription file should have identical number of lines. You
need to repeat current transcription same number of times as many speakers are
in your database.
thanks so very much for your help. because of you, I have a shot at passing my
thesis. thank you!