You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We currently do not have an effective way to build a homogeneous or heterogeneous graph from graph database data (i.e. a set of files with non-contiguous vertex and edge ids labeled by vertex and edge type). We should create a means of doing this using PyTorch/NCCL and add appropriate examples. In the long term, it might make sense to contribute this to upstream PyG.
We should also provide a function that handles when there is an imbalance of labeled data across ranks by evening out the label distribution through shuffling.
The text was updated successfully, but these errors were encountered:
We currently do not have an effective way to build a homogeneous or heterogeneous graph from graph database data (i.e. a set of files with non-contiguous vertex and edge ids labeled by vertex and edge type). We should create a means of doing this using PyTorch/NCCL and add appropriate examples. In the long term, it might make sense to contribute this to upstream PyG.
We should also provide a function that handles when there is an imbalance of labeled data across ranks by evening out the label distribution through shuffling.
The text was updated successfully, but these errors were encountered: