The problem to use rbm-cf of GraphChi

User 1205 | 1/15/2015, 9:33:52 AM


I meet a problem when I use rbm-cf to deal with the dataset--smallnetflix_mm.

My instructs are as follows: mc@mc-virtual-machine:~/Downloads/graphchi-cpp-master$ ./toolkits/collaborativefiltering/rbm --training=smallnetflixmm --validation=smallnetflixmme --minval=1 --maxval=5 --maxiter=6 --quiet=0 –D=100 –rbmalpha=0.01 rbmbeta=0.02

The results are as follows: WARNING: common.hpp(print_copyright:214): GraphChi Collaborative filtering library is written by Danny Bickson (c). Send any comments or bug reports to [training] => [smallnetflix_mm] [validation] => [smallnetflixmme] [minval] => [1] [maxval] => [5] [max_iter] => [6] [quiet] => [0] INFO: sharder.hpp(startpreprocessing:370): Starting preprocessing, shovel size: 17476266 Segmentation fault (core dumped)

I don't know what's the problem is. What dose it mean by "Segmentation fault (core dumped)"?

Thanks, Allen


User 6 | 1/15/2015, 6:31:32 PM

Hi 1) Please note that you have mistake in the command line parameters starting from -D instead of --D the same with rbmalpha and rbmbeta 2) It seems you are running in a virtual machine. Check that you have enough cores (at least 2) enough memory for the virtual machine (at least 4GB) and the virtual machine is 64 bit and not 32 bit os. 3) If everything above confirms to the recommendations please compile with "make clean; make cfd" and send me the full output

User 1205 | 1/16/2015, 12:59:28 AM

Hi, Danny

Thank you for your reply. Now I am trying to change my virtual machine into 64 bit.