[MINC-users] best_nonlin_reg.pl ?

[Lisa F. Akiyama]

Hi Andrew,

Thanks for uploading your copy of bestlinreg and suggesting an alternative
way (using -linmethod 'minctracc -lsq9').

I got rid of the previous errors I was getting, but now I get the following:

Initial objective function val = 0.16510932
> Final objective function value = 0.14109196
> mv -f /tmp/bestlinreg-SMx3R3ux/Sfit-model-iso_4.xfm
> genmodel-work/00/0003-SEA0094_6mo_May2009_t1mpr.xfm
> Died at /usr/local/bic/bin/xfmavg line 310.
> Died at /usr/local/bic/bin/volgenmodel line 656.


I get the above when using your copy of bestlinreg, a copy of bestlinreg
that Claude had provided me, and even when running with the -linmethod
'minctracc -lsq9' option.
The output file that I specified the name of is no where to be found in the
workdir or the dir where I have stored my input files.

Any tips?

Many thanks.


Best,
Lisa



On Tue, Mar 1, 2011 at 5:22 PM, Andrew Janke <a.janke at gmail.com> wrote:

> > Yiiikes. One more script (hopefully this will be the last; I really need
> to
> > stop bugging you) that I need... : bestlinreg.
>
> Ah, well this one is Claudes. It's a condensed version of mritotal
> using the same code layout as nlpfit.
>
> I have uploaded the version I have of it to the scripts directory but
> note that it may not be the latest or best version. Claude will no
> doubt correct me.
>
> Another way around this is if you know that your brains are all fairly
> close to each other is to do this:
>
>   $ volgenmodel -linmethod 'minctracc -lsq9' ....
>
> I should also mention that once you have it all working, if you have a
> gridengine install add -batch argument and watch the fun begin as it
> brings your network/cluster to a crawl.
>
>
> a
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>