train ivector

[wjyfelicity]

Recently, I am doing speaker diarization with LIUM, After I trained Total variability matrix and i-vector, when train EFR normalization, I met some problem about Eigenvalue Decomposition failed, because org.ejml.simple.simpleEVD.(Unknown Source).Do you know what happened about it.

[StevenLOL]

Hi,

I am sorry, I am not familiar with ivector training in LIUM. You may contact LIUM team for this issue.

We usually train ivector system with Kaldi, eg SRE2010 and SRE2016.

[glynpu]

Hi,

I am trying i-vector ilp clutering with LIUM and UBM has been trained with Alize toolkit and import by LIUM successfully. However, I am stuck in total variability training.

I think lots of recordings will be needed to train a TV matrix, just like the UBM training. According to training script TV.sh http://www-lium.univ-lemans.fr/diarization/doku.php/i-vector , the command seems only take a single mfcc file as training features, how should I input lots of recordings input the command? It seems that all feature should be put into a single .mfcc file?

segIn=./seg/ubm.seg
gmm=./ubm/ubm.gmm
fMask=./mfcc/%s.mfcc

[StevenLOL]

Hi, @glynpu

fMask=./mfcc/%s.mfcc

Sorry for the late reply, you may already figure it out, this fMask means all mfcc files in ./mfcc/ folder.