[ Y ] Have you updated to latest MFA version?
tried on 3.0.0a and stable 2.2.17
[ Y ] Have you tried rerunning the command with the --clean flag?
Describe the issue
The full corpus contains ~100 files, and 80% of the alignment was successful.
The 20% unsuccessful files was not warned anywhere in the cmd, need to manually locate them.
Tried to isolate the issue of some files, then KaldiProcessingError happened.
Single file would just output KaldiProcessingError, nothing in output folder
Multiple files also didn't work, tried with 7 files where 1 was unsuccessful, and the error occured, none of the previously successful files were outputed
For Reproducing your issue
Please fill out the following:
Corpus structure
What language is the corpus in? Cantonese
How many files/speakers? ~100 files
Are you using lab files or TextGrid files for input? .wav with .txt files
Dictionary
Are you using a dictionary from MFA? If so, which one? no
If it's a custom dictionary, what is the phoneset? Jyutping
Acoustic model
If you're using an acoustic model, is it one download through MFA? If so, which one? no, it was found online from previous studies
If it's a model you've trained, what data was it trained on? no
Log file
Please attach the log file for the run that encountered an error (by default these will be stored in ~/Documents/MFA).
Traceback (most recent call last):
File "C:\Users\xxx\anaconda3\envs\aligner\lib\site-packages\montreal_forced_aligner\abc.py", line 92, in run
yield from self._run()
File "C:\Users\xxx\anaconda3\envs\aligner\lib\site-packages\montreal_forced_aligner\alignment\multiprocessing.py", line 1003, in _run
self.check_call(align_proc)
File "C:\Users\xxx\anaconda3\envs\aligner\lib\site-packages\montreal_forced_aligner\abc.py", line 119, in check_call
raise KaldiProcessingError([self.log_path])
LOG (gmm-boost-silence.EXE[5.5.1068]:main():gmmbin\gmm-boost-silence.cc:93) Boosted weights for 5 pdfs, by factor of 1
LOG (gmm-boost-silence.EXE[5.5.1068]:main():gmmbin\gmm-boost-silence.cc:103) Wrote model to -
add-deltas 'scp,s,cs:C:\Users\xxx\Documents\MFA\tempFolder\tempFolder\split8\feats.1.1.scp' ark:-
LOG (gmm-align-compiled.EXE[5.5.1068]:main():gmmbin\gmm-align-compiled.cc:127) 1-1
WARNING (gmm-align-compiled.EXE[5.5.1068]:kaldi::AlignUtteranceWrapper():decoder\decoder-wrappers.cc:617) Retrying utterance 1-1 with beam 40
WARNING (gmm-align-compiled.EXE[5.5.1068]:kaldi::AlignUtteranceWrapper():decoder\decoder-wrappers.cc:626) Did not successfully decode file 1-1, len = 11423
LOG (gmm-align-compiled.EXE[5.5.1068]:main():gmmbin\gmm-align-compiled.cc:135) Overall log-likelihood per frame is -nan(ind) over 0 frames.
LOG (gmm-align-compiled.EXE[5.5.1068]:main():gmmbin\gmm-align-compiled.cc:137) Retried 1 out of 1 utterances.
LOG (gmm-align-compiled.EXE[5.5.1068]:main():gmmbin\gmm-align-compiled.cc:139) Done 0, errors on 1
Desktop (please complete the following information):
OS: Windows
Version Windows 11
Any other details about the setup (Cloud, Docker, etc)
Anaconda
Additional context
After changing the --beam 1000 --retry-beam 4000, all files was aligned, but the runtime was much longer.
IDK why, but the --finetune flag couldn't output any textgrid too, even tho the first and second pass alignment were successful
Debugging checklist
[ Y ] Have you updated to latest MFA version? tried on
3.0.0a
and stable2.2.17
[ Y ] Have you tried rerunning the command with the--clean
flag?Describe the issue The full corpus contains ~100 files, and 80% of the alignment was successful.
Tried to isolate the issue of some files, then
KaldiProcessingError
happened.KaldiProcessingError
, nothing in output folderFor Reproducing your issue Please fill out the following:
Cantonese
~100 files
.wav with .txt files
no
Jyutping
no, it was found online from previous studies
no
Log file Please attach the log file for the run that encountered an error (by default these will be stored in
~/Documents/MFA
).Desktop (please complete the following information):
Windows
Windows 11
Anaconda
Additional context After changing the
--beam 1000 --retry-beam 4000
, all files was aligned, but the runtime was much longer. IDK why, but the--finetune
flag couldn't output any textgrid too, even tho the first and second pass alignment were successful