graph2mat.bindings.e3nn.modules.edge_operations

E3nn operations to compute edge matrix blocks.

In edge matrix blocks, you tipically will have, for each edge, a different message coming from each atom in the edge. The edge block will tipically not be symmetric, but it is common that.

\[B_{ij} = B_{ji}^T\]

Classes

E3nnEdgeBlockNodeMix(edge_feats_irreps, ...)
E3nnSimpleEdgeBlock(irreps_in, irreps_out)

class graph2mat.bindings.e3nn.modules.edge_operations.E3nnEdgeBlockNodeMix(edge_feats_irreps: Irreps, edge_messages_irreps: Irreps, node_feats_irreps: Irreps, irreps_out: Irreps)

Bases: Module

__init__(edge_feats_irreps: Irreps, edge_messages_irreps: Irreps, node_feats_irreps: Irreps, irreps_out: Irreps)

Initialize internal Module state, shared by both nn.Module and ScriptModule.

forward(edge_feats: Tuple[Tensor, Tensor], edge_messages: Tuple[Tensor, Tensor], node_feats: Tuple[Tensor, Tensor]) → Tensor

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class graph2mat.bindings.e3nn.modules.edge_operations.E3nnSimpleEdgeBlock(irreps_in: Irreps, irreps_out: Irreps)

Bases: Module

__init__(irreps_in: Irreps, irreps_out: Irreps)

Initialize internal Module state, shared by both nn.Module and ScriptModule.

forward(**tuple_kwargs: Tuple[Tensor, Tensor]) → Tensor

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.