search

Thinklab-SJTU / ThinkMatch

A research protocol for deep graph matching.
https://thinkmatch.readthedocs.io/
Other
826 stars 118 forks source link

Can GANN-MGM handle large graph ? #48

Closed xiachenrui closed 2 years ago

xiachenrui commented 2 years ago

Hi, your work is really great! I would like to know if the GANN-MGM model can work on very large graphs (with about 30,000 nodes and features of each node are known). Do you have any other suggestions about work on large graph?

Thanks a lot !

rogerwwww commented 2 years ago

Hi Chenrui,

If your graph scales up to 30,000 nodes, perhaps you should first try two-graph matching methods that are more memory efficient. For the GANN-MGM model, the GPU memory might be an issue (or perhaps you could try using CPU and RAM).

I feel like you might be interested in our new pygmtools: https://github.com/Thinklab-SJTU/pygmtools It provides numpy&pytorch implementations of graph matching methods with a unified python interface.

xiachenrui commented 2 years ago

Thanks for reply! I can't wait to try it.