search

leoribeiro / struc2vec

This repository provides a reference implementation of struc2vec.
MIT License
381 stars 116 forks source link

The problem about the scalability #8

Open Tomposon opened 6 years ago

Tomposon commented 6 years ago

I tried to use the struc2vec to train on 1 million nodes with 24 threads,but no embeddings was generated after three days.Is there a solution?

xinliweiyuan commented 5 years ago

Have you solved the problem? I have the same need on the big graphs.

leoribeiro commented 5 years ago

Have you tried to turn on the optimizations?

ShengdingHu commented 5 years ago

1 million nodes may be too large for struc2vec. For me. a 80 ,000 nodes graph with threads 24 hasn't generate embedding after 5 hours training. And the algorithm is O(n^2logn) , so may be will take you about 200 *5 + hours.

ShengdingHu commented 5 years ago

But I am wondering that, is there an upper bound for the number of threads we use? i.e. can the algorithm parallel well using large number of threads? Have the author test it? Thank you very much

leoribeiro commented 5 years ago

@ShengdingHu Yes. The algorithm works with a large number of threads. However, it can consume much memory.