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
The official SALIENT++ system described in the paper "Communication-Efficient Graph Neural Networks with Probabilistic Neighborhood Expansion Analysis and Caching".
Version of pytorch-metis that has bindings that directly accept pytorch tensors for constructing metis graphs. No data copies are required to convert from pytorch CSR and Metis's graph representation because METIS uses CSR internally.