Montreal-Forced-Aligner
Montreal-Forced-Aligner copied to clipboard
MontrealCorpusTools
when I used the command like "mfa validate ~/mfa_data/my_corpus ~/mfa_data/my_dictionary.txt",I met some erors:KaldiProcessingError. My audio is 7 minutes long
/data2/zmy/kaldi/src/gmmbin/gmm-boost-silence --boost=1.25 1 /home/zmy/Documents/MFA/align_validate_training/monophone/1.mdl - WARNING (gmm-boost-silence[5.5.1035~2-3dd90]:main():gmm-boost-silence.cc:82) The pdfs for the silence phones may be shared by other phones (note: this probably does not matter.) LOG (gmm-boost-silence[5.5.1035~2-3dd90]:main():gmm-boost-silence.cc:93) Boosted weights for 5 pdfs, by factor of 1.25 /data2/zmy/kaldi/src/gmmbin/gmm-align-compiled --transition-scale=1.0 --acoustic-scale=0.1 --self-loop-scale=0.1 --beam=6 --retry-beam=40 --careful=false - ark:/home/zmy/Documents/MFA/align_validate_training/monophone/fsts.1.0.ark 'ark,s,cs:apply-cmvn --utt2spk=ark:"/home/zmy/Documents/MFA/align_validate_training/align/split1/utt2spk.1.0.scp" scp:"/home/zmy/Documents/MFA/align_validate_training/align/split1/cmvn.1.0.scp" scp:"/home/zmy/Documents/MFA/align_validate_training/align/split1/feats.1.0.scp" ark:- | add-deltas ark:- ark:- |' ark:/home/zmy/Documents/MFA/align_validate_training/monophone/ali.1.0.ark ark,t:- LOG (gmm-boost-silence[5.5.1035~2-3dd90]:main():gmm-boost-silence.cc:103) Wrote model to - add-deltas ark:- ark:- apply-cmvn --utt2spk=ark:/home/zmy/Documents/MFA/align_validate_training/align/split1/utt2spk.1.0.scp scp:/home/zmy/Documents/MFA/align_validate_training/align/split1/cmvn.1.0.scp scp:/home/zmy/Documents/MFA/align_validate_training/align/split1/feats.1.0.scp ark:- LOG (apply-cmvn[5.5.1035~2-3dd90]:main():apply-cmvn.cc:162) Applied cepstral mean normalization to 1 utterances, errors on 0 LOG (gmm-align-compiled[5.5.1035~2-3dd90]:main():gmm-align-compiled.cc:127) 1-1 WARNING (gmm-align-compiled[5.5.1035~2-3dd90]:AlignUtteranceWrapper():decoder-wrappers.cc:617) Retrying utterance 1-1 with beam 40 WARNING (gmm-align-compiled[5.5.1035~2-3dd90]:AlignUtteranceWrapper():decoder-wrappers.cc:626) Did not successfully decode file 1-1, len = 44998 LOG (gmm-align-compiled[5.5.1035~2-3dd90]:main():gmm-align-compiled.cc:135) Overall log-likelihood per frame is -nan over 0 frames. LOG (gmm-align-compiled[5.5.1035~2-3dd90]:main():gmm-align-compiled.cc:137) Retried 1 out of 1 utterances. LOG (gmm-align-compiled[5.5.1035~2-3dd90]:main():gmm-align-compiled.cc:139) Done 0, errors on 1
INFO - Corpus INFO - ******
INFO - 1 sound files INFO - 1 text files INFO - 1 speakers INFO - 1 utterances INFO - 450.000 seconds total duration
INFO - Sound file read errors INFO - ======================
INFO - There were no issues reading sound files.
INFO - Feature generation INFO - ==================
INFO - There were no utterances missing features.
INFO - Files without transcriptions INFO - ============================
INFO - There were no sound files missing transcriptions.
INFO - Transcriptions without sound files INFO - ==================================
INFO - There were no transcription files missing sound files.
INFO - ********** INFO - Dictionary INFO - **********
INFO - Out of vocabulary words INFO - =======================
INFO - 29 OOV word types INFO - 403 total OOV tokens
INFO - For a full list of the word types, please see:
INFO - /home/zmy/Documents/MFA/align_validate_training/oovs_found.txt
INFO - For a by-utterance breakdown of missing words, see:
INFO - /home/zmy/Documents/MFA/align_validate_training/utterance_oovs.txt
INFO - ******** INFO - Training INFO - ********
INFO - Initializing training for monophone... INFO - Compiling training graphs... 100%|██████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:01<00:00, 1.28s/it] INFO - Generating initial alignments... 100%|██████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:01<00:00, 1.27s/it] INFO - monophone training already initialized, skipping initialization. INFO - monophone - Iteration 1 of 40 INFO - Generating alignments... 0%| | 0/1 [00:22<?, ?it/s] ERROR - There was an error in the run, please see the log. MultiprocessingError:
Job 0 encountered an error: Traceback (most recent call last):
File "/home/zmy/anaconda3/envs/aligner/lib/python3.8/site-packages/montreal_forced_aligner/abc.py", line 73, in run
yield from self._run()
File
"/home/zmy/anaconda3/envs/aligner/lib/python3.8/site-packages/montreal_forced_aligner/alignment/multiprocessing.py",
line 503, in _run
self.check_call(align_proc)
File "/home/zmy/anaconda3/envs/aligner/lib/python3.8/site-packages/montreal_forced_aligner/abc.py", line 100, in
check_call
raise KaldiProcessingError([self.log_path])
montreal_forced_aligner.exceptions.KaldiProcessingError: KaldiProcessingError:
There were 1 job(s) with errors when running Kaldi binaries.
See the log files below for more information.
/home/zmy/Documents/MFA/align_validate_training/monophone/log/align.1.0.log
here is my error,thank you!
You can try increasing the beam in the mfa align ... --beam 100 and see if it aligns? The default settings are geared towards processing lots of shorter files rather than long files with no extra time information. Any ways you can cut down the audio into smaller chunks will go a long way for generating accurate alignments.
