ValueError: 'diag' covars must be positive

[Obile19]

Hi, When I pass 0 or bad number of speaker to diarization function, I have an error! What should I do now?

[oulfik]

Have the same issue when setting numberOfSpeakers to 0 while using the command-line example in the wiki.

C:\Users\oliver\Desktop\digMed\Datenstrukturen und Analysemethoden im Gesundheitsbereich\diarization\pyaudioanalysis\pyAudioAnalysis\pyAudioAnalysis\audioSegmentation.py:326: RuntimeWarning: invalid value encountered in true_divide transmutation_matrix[i, :] /= transmutation_matrix[i, :].sum() Traceback (most recent call last): File "C:\Users\oliver\Desktop\digMed\Datenstrukturen und Analysemethoden im Gesundheitsbereich\diarization\pyaudioanalysis\pyAudioAnalysis\pyAudioAnalysis\audioAnalysis.py", line 617, in speakerDiarizationWrapper(args.input, args.num, args.flsd) File "C:\Users\oliver\Desktop\digMed\Datenstrukturen und Analysemethoden im Gesundheitsbereich\diarization\pyaudioanalysis\pyAudioAnalysis\pyAudioAnalysis\audioAnalysis.py", line 239, in speakerDiarizationWrapper aS.speaker_diarization(inputFile, numSpeakers, lda_dim=0, plot_res=True) File "C:\Users\oliver\Desktop\digMed\Datenstrukturen und Analysemethoden im Gesundheitsbereich\diarization\pyaudioanalysis\pyAudioAnalysis\pyAudioAnalysis\audioSegmentation.py", line 1005, in speakerdiarization hmm.means = means; hmm.covars = cov File "C:\Users\oliver\Desktop\digMed\Datenstrukturen und Analysemethoden im Gesundheitsbereich\diarization\pyaudioanalysis\pyAudio-env\lib\site-packages\hmmlearn\hmm.py", line 182, in covars _utils._validate_covars(covars, self.covariance_type, File "C:\Users\oliver\Desktop\digMed\Datenstrukturen und Analysemethoden im Gesundheitsbereich\diarization\pyaudioanalysis\pyAudio-env\lib\site-packages\hmmlearn_utils.py", line 28, in _validate_covars raise ValueError("'diag' covars must be positive") ValueError: 'diag' covars must be positive