search

razvancaramalau / Sequential-GCN-for-Active-Learning

50 stars 17 forks source link

Super adjacency matrix #8

Closed lijiaoyang closed 1 year ago

lijiaoyang commented 2 years ago

For large data sets such as Imagenet, when constructing adjacency matrix, the program memory and video memory will explode. Is there any good solution

razvancaramalau commented 2 years ago

This is a known limitation for GCNs in general. However, there are a few directions to mitigate this issue. The one mentioned in the paper is by creating a random subset of unlabelled < entire unlabelled pool. Another way is to sub-graph the feature space according to groups of your labelled data. (This one I also had it implemented and it works). Finally, you can look into graph coarsening. There are some more elegant solutions in there.

lijiaoyang commented 2 years ago

thanks for your reply!! about "Another way is to sub-graph the feature space according to groups of your labelled data. ", could you be more specific? Or any example code？

thanks again !!