GNNs for small networks #3768
-
|
Hello. Sorry if this is a bit of a general question, but I don't see a lot of interest in it. I think there is a focus on making GNNs for large graphs, but I'm interested in a problem that has millions of small graphs, less than 20 nodes each. Is there a benchmark dataset for such a problem and are there ideal GNNs in pytorch geometric? I'm looking into the research so far but I'm a bit stuck. Thanks. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 2 replies
-
|
PyG is well-suited for these kinds of problems in my opinion. For example, we have a variety of benchmark examples for all OGB graph-level tasks here. I think the |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for the quick reply. Is there some advantage with the GIN model or is it just a bunch of GINConvs as in the example? |
Beta Was this translation helpful? Give feedback.
-
|
No, they are the same. We refer to GIN as a stack of |
Beta Was this translation helpful? Give feedback.
PyG is well-suited for these kinds of problems in my opinion. For example, we have a variety of benchmark examples for all OGB graph-level tasks here. I think the
GINnetwork is a great starting place.