vedaldi / svm-struct-matlab Goto Github PK

When {patterns} are sparse vectors with high dimension, MATLAB crashes with memory issue because model.w is not a sparse vector but dense. I don't know how to make it sparse...

hello,thank you for your work. I'm confused how to use this to deal with classify problem. I'm using windows system. maybe it haven't compile the 'svm_struct_classify.c' in 'makefile_windows.m' file. hope for your reply,thank you.

hello
i am receiving this error while running the makefile_windows command:
"Deprecated MEX function mexSetTrapFlag was called. Update the source code to use mexCallMATLABWithTrap."

i would appreciate any advice on how to solve this

BR

Noam

Hello!

I've been encountered few errors when I executing makefile_windows which as you can see below:

doing hideo
Building with 'Microsoft Visual C++ 2013 Professional (C)'.
Error using mex
svm_hideo.c
c:\users\ibrahimpc2\documents\matlab\svm-struct-matlab-master\svm_light\svm_common.h(447) : error C2059: syntax error : 'sizeof'

Error in makefile_windows (line 3)
mex -largeArrayDims -c -DWIN ./svm_light/svm_hideo.c

Can somebody guide me on how to solve it?

/home/sattar/Downloads/svm-struct-matlab-1.2/svm_struct_learn_mex.c:108:3: error: expected expression before '/' token
/home/sattar/Downloads/svm-struct-matlab-1.2/svm_struct_learn_mex.c:188:3: error: expected expression before '/' token
I have this problem . I change the command line with /* */ but it is not wokring for me.

Hi,
I'm working in Linux Fedora 18 with MATLAB R2011b and GCC 4.6. When I tried to compile the code I got this error:

svm_struct_learn_mex.c: In function ‘mexFunction’:
svm_struct_learn_mex.c:108:3: error: expected expression before ‘/’ token
svm_struct_learn_mex.c:188:3: error: expected expression before ‘/’ token

mex: compile of ' "svm_struct_learn_mex.c"' failed.

I replaced the comment in the line 108 and 188 with
/* .... */

Currently, I can compile the mex-files succesfully.

The name is slightly confusing. It is possible to rename the repository through Repository Administration panel.

Hello,

Is it possible to initialize the model weights with some prior values?

Thank you
Yuval