Thanks, Steve. I can’t find the old data upload URL, so here’s a Dropbox link to a FIX_debugging.tar.gz file containing all of the output and error files from my latest FIX training attempt. 

https://www.dropbox.com/s/h1o6uuvni72anvr/FIX_debugging.tar.gz

I’ve copied individual features (from the subject.feat/fix folder) to the “training_features” subfolder. 

Thanks for taking the time to look at this.

Cheers,

David

On Nov 8, 2014, at 4:21 AM, Stephen Smith <[log in to unmask]> wrote:

Hi David - it's hard to tell from this what's going wrong with the training-weights creation. If you want to create a tarfile containing all of the output folders/files being created we can take a look.

Regards, Steve.

On 7 Nov 2014, at 19:42, David V. Smith <[log in to unmask]> wrote:

Hi Steve,

My mistake — I wasn’t actually using the compiled files (I didn’t change the default). FSL_FIX_MATLAB_MODE=1

Unfortunately, I don’t have SGE set up here; I’m not sure that option is set up by default (it doesn’t seem to be). 

I tried FIX with the Standard.RData training weight files, and it worked fine (as did the subsequent cleaning step). So, it looks like I’m mainly having trouble generating a study-specific <training>.RData file. Any ideas?

Thanks for your help. This is really cool software, and I’m excited to use it!

Cheers,

David

On Nov 7, 2014, at 2:23 AM, Stephen Smith <[log in to unmask]> wrote:

Hi - I would not necessarily recommend that you recompile the matlab - that shouldn't be necessary?

The LOO option will use SGE if it's setup - so you may want to check on that - and also try running a single FIX run with a presupplied training-weights file to see if that runs OK.

Cheers.

On 6 Nov 2014, at 21:28, David V. Smith <[log in to unmask]> wrote:

Hi,

I’m also having trouble setting up FIX. Our sysadmin has compiled the Matlab binaries and has changed the paths in the settings.sh file (described in the README file).

When I train the classifier on my hand labels without the LOO, there’s no training.Rdata output file. And when I use the LOO option during training, it produces the errors below and dumps several files with strange names (e.g., "302:.e6151" and "302:.o6151”):

leaving-out: /mnt/delgadolab/disk02/labdata02/AffInf01/FSL/203/run1.feat
/bin/sh: 1: 302:: not found
/bin/sh: 1: 302:: not found
/bin/sh: 1: 302:: not found
/bin/sh: 1: 302:: not found
/bin/sh: 1: 302:: not found
/bin/sh: 1: 302:: not found
/bin/sh: 1: 302:: not found
/bin/sh: 1: 302:: not found

I’m not sure how to debug from here. It looks like our environment is set up correctly when I source the settings.sh file. The only other I error I see is in one of the matlab logs:

tail: cannot open `/mnt/delgadolab/disk02/labdata02/AffInf01/FSL/203/run1.feat/fix4melview_training3_LOO_thr1.txt' for reading: No such file or directory

But I’m assuming that file didn’t get created because something broke earlier…

Thanks for any help!

Cheers,
David



On Nov 2, 2014, at 7:26 AM, Alireza Sedghi <[log in to unmask]> wrote:

Dear FSL experts
i want to install FIX on my ubuntu 14.04 , i have downloaded it and unpacked it to my home directory , the setting.sh is somehow confusing ! i have MATLAB installed already , which part of setting.sh must change ?
how can i change the matlab startup option
i really appreciate if you can help

----------------------------------------------
David V. Smith, Ph.D.
Postdoctoral Fellow, Delgado Lab
Department of Psychology
Rutgers University
Newark, NJ 07102
----------------------------------------------

---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director,  Oxford University FMRIB Centre

FMRIB, JR Hospital, Headington, Oxford  OX3 9DU, UK
+44 (0) 1865 222726  (fax 222717)
[log in to unmask]    http://www.fmrib.ox.ac.uk/~steve

---------------------------------------------------------------------------

Stop the cultural destruction of Tibet

----------------------------------------------

David V. Smith, Ph.D.
Postdoctoral Fellow, Delgado Lab

Department of Psychology

Rutgers University

Newark, NJ 07102

----------------------------------------------

---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director,  Oxford University FMRIB Centre

FMRIB, JR Hospital, Headington, Oxford  OX3 9DU, UK
+44 (0) 1865 222726  (fax 222717)
[log in to unmask]    http://www.fmrib.ox.ac.uk/~steve

---------------------------------------------------------------------------

Stop the cultural destruction of Tibet