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
Trying to dive into how sparse-sparse matrix multiplication is working but I can't figure out where this import comes from, if I try importing it in a notebook with torch and torch_sparse imported I get an import error, but I see it's used in the source code of matmul.py. Where does this come from?
The text was updated successfully, but these errors were encountered:
The implementation of spspmm is here (CPU) and here (GPU). For GPU, we make use of cusparse routines, while we implement our own simple variant for CPU-processing.
Trying to dive into how sparse-sparse matrix multiplication is working but I can't figure out where this import comes from, if I try importing it in a notebook with
torch
andtorch_sparse
imported I get an import error, but I see it's used in the source code ofmatmul.py
. Where does this come from?The text was updated successfully, but these errors were encountered: