Closed hyx321 closed 6 years ago
@hyx321 Thanks for your interests in our work.
What do you mean by nothing? Can you give more information about it, like screenshots? That will be very helpful.
After I run the code :"python 200k_commu_mat_computation.py yelp-50k all 1",I got a error.And I also conda install bottleneck.I use windows OS.
@hyx321
What's your version of bottleneck? Mine is
bottleneck.__version__
Out[3]:
'1.1.0'
Seems the method argpartsort is not available in your bottleneck.
today ,follow your advice,I changed the version.And the error was gone.but I get another error.By the way,I revised the code that is "f = open(ind2rid_filename, 'rb')",this code is at 265lines in 200k_commu_mat_computation.py .Tomrrow ,I'lltry to run this Project in Linux.
today,I build a virtual machine of Ubuntu.After run this project ,the last error about"cpickle.UnpicklineError" has gone and I get a sad error about Memory,I give it about 6.5G Memory.When it run ,My windows only have 1G for free .and I have 12G in total .
@hyx321 I test the whole program again, I think you need around 16G Memory in Linux server.
Thank you for your helping these days.After receiving your e-mail , I try to decrease datas properly in every txt ,and I get a error " Keyerror :640698".So ,do you have a smaller data to let me run this program?please.
@hyx321 If what you need is to execute the whole process, you may comment the Line 458-460, which caused the memory problem. It computed two meta-graph based similarity matrices, however, you can execute the whole process without them. The needed latent features are already in the directory "yelp-50k/exp_split/1/mf_features/path_count".
Thank you for you advice ,I run this program successful,and it only need 6.1G.Honestly,It is a good news for me. Parsing code is my next step .
this week ,I get nothing when I open the website of readme several times.md.It shows nothing .