Hi, I'm using this framework for a demo project of GCN acceleration, but it seems that operations between a sparse matrix of COO format (e.g. edge_index compressed) and a matrix of dense format (e.g. intermediate result of feature matrix and weights) is not supported.
Hi, I'm using this framework for a demo project of GCN acceleration, but it seems that operations between a sparse matrix of COO format (e.g. edge_index compressed) and a matrix of dense format (e.g. intermediate result of feature matrix and weights) is not supported.