Batching with multiple graphs #8180
-
|
I have a dataset in which each example is a list of heterogeneous graphs associated with a numeric label, e.g. X = [graph1, graph2, graph3, graph4, graph5] and y = 1.2345. I would like to use the PyG dataloader to feed them into a model, but obviously this requires loading multiple graphs per item The Advanced Mini-Batching article lays out how to do this for a pair of graphs, but not more than this. Is this possible to do, and, if so, how should I go about it? Thanks! |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
Your use case sounds very similar to #7928. Have you had a chance to try #7928 (comment)? |
Beta Was this translation helpful? Give feedback.
-
|
Apologies, I searched through the previous comments but didn't find that one. Thanks for directing me to it--yes, it solves my problem |
Beta Was this translation helpful? Give feedback.
Your use case sounds very similar to #7928. Have you had a chance to try #7928 (comment)?