bimodal GMM training crashes/freezes

[chaosct]

I am using the python bindings of XMM, compiled in Ubuntu 15.10 (see #2).

More often than not, the GMM.train() method crashes/freezes when using bimodal GMM (regression). It has never crashed with GMM classification so far.

Example script: https://gist.github.com/chaosct/71e9f1412397f6e67d33

Executing this example sometimes works, sometimes freezes (and never returns), and sometimes crashes with:

python: ../nptl/pthread_mutex_lock.c:355: __pthread_mutex_lock_full: Assertion `(-(e)) != 3 || !robust' failed.

[JulesFrancoise]

Hi,

Thank you for the bug report. I tested your example script but I cannot reproduce the bug on Mac OS X. I will try to reproduce it on a linux machine as soon as possible.

Thanks,
Jules