dgl.ops.e_sub_vΒΆ

dgl.ops.e_sub_v(g, x, y)ΒΆ

Generalized SDDMM function. It computes edge features by sub edge features and destination node features.

Parameters
  • g (DGLGraph) – The input graph

  • x (tensor) – The edge features.

  • y (tensor) – The destination node features.

Returns

The result tensor.

Return type

tensor

Notes

This function supports autograd (computing input gradients given the output gradient). If the feature shape of two input operands do not match, we first broadcasts the features to a unified shape (note that the memory usage will not increase accordingly) and then performs the operation.

Broadcasting follows NumPy semantics. Please see https://docs.scipy.org/doc/numpy/user/basics.broadcasting.html for more details about the NumPy broadcasting semantics.