search

StochasticNumerics / mimclib

A software library for UQ methods
GNU General Public License v2.0
6 stars 8 forks source link

NaN in Variance #96

Closed litvinen closed 7 years ago

litvinen commented 7 years ago

I execute ./echo_test_cmd_new.py -tries 5 | parallel -j5

and obtain mimc.py:664: (RuntimeWarning) divide by zero encountered in power mimc.py:667: (RuntimeWarning) invalid value encountered in multiply mimc.py:668: (RuntimeWarning) invalid value encountered in multiply fromnumeric.py:2645: (VisibleDeprecationWarning) rank is deprecated; use the ndim attribute or function instead. To find the rank of a matrix see numpy.linalg.matrix_rank.

TOL 0.08 Time since last tic: 0.0233 sec. Doing 10 of level [0] Doing 10 of level [1] Doing 10 of level [2] theta nan New M: [1 1 1] Time since last tic: 0.2881 sec. Eg=1.06769967218 Bias=nan StatErr=nan TotalErrEst=nan | 8.000000000000e-02 Level E V sampleV M Time [0] +1.282385197338e+00 6.241426827345e-04 6.241426827345e-04 10 2.540801e-02 [1] +1.371630417309e-01 nan 2.885077937117e-04 10 1.325130e-04 [2] +7.752248343102e-02 nan 5.125359292815e-05 10 2.211094e-04

Time since last tic: 0.2895 sec. MIMC iteration for TOL=0.08 took 0.351119995117 seconds SEE PLEASE THE NEXT POST

litvinen commented 7 years ago

If I put -mimc_bayesian False, I do not receive NaN and this error in mimc.py

litvinen commented 7 years ago

I think the problem is solved. The reason was that coefficient beta=1 (is wrong, correct in my case is 2)