ALIZE-Speaker-Recognition / LIA_RAL

A high-level toolkit for speaker recognition, build on top of ALIZE-Core.
http://alize.univ-avignon.fr
GNU Lesser General Public License v3.0
125 stars 27 forks source link

I use the 01_GMM-UBM_system_with_ALIZE3.0 example #1

Closed sunjunlishi closed 7 years ago

sunjunlishi commented 7 years ago

EnergyDetector --config cfg/EnergyDetectorSPro.cfg --inputFeatureFilename data/data.lst --featureFilesPath data/prm/ --labelFilesPath data/lbl/ i have aproblem that xg22 s0e2b5ctri0qy7005 ![mjptem6rl7dk cvd 6t](https://cloud.githubusercontent.com/assets/11845664/19913524/526e8f02-a0de-11e6-9ba8-ef1981e75d55.png)

jfb84 commented 7 years ago

Hello

By using verboseLevel and/or debug options, and by redirecting the standard output into a file, you could send a text with the syssteem outputs. Easier... We need also the description of the problem and the configuration file of EnergyDetector.

It seems to me that there is no model adaptation. Look at the options to adapt means, weights and cov (meanAdapt==true ...)

Best JF

De: "sunjunlishi" notifications@github.com À: "ALIZE-Speaker-Recognition/LIA_RAL" LIA_RAL@noreply.github.com Envoyé: Mercredi 2 Novembre 2016 02:25:54 Objet: [ALIZE-Speaker-Recognition/LIA_RAL] I use the 01_GMM-UBM_system_with_ALIZE3.0 example (#1)

EnergyDetector --config cfg/EnergyDetector_SPro.cfg --inputFeatureFilename data/data.lst --featureFilesPath data/prm/ --labelFilesPath data/lbl/ i have aproblem that

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub , or mute the thread .


Jean-Francois BONASTRE Directeur du LIA LIA/CERI Université d'Avignon Tel: +33/0 490843514 directeur-lia@univ-avignon.fr


sunjunlishi commented 7 years ago

oh,it works,its my problem;I change the 'sfbcep.exe',it works,thanks. 3 q 8t1_ wr6vt 2 76 g e