gnn_tracking.models.dynamic_edge_conv

gnn_tracking.models.dynamic_edge_conv#

Attributes#

knn

Classes#

DynamicEdgeConv

Module Contents#

gnn_tracking.models.dynamic_edge_conv.knn = None#
class gnn_tracking.models.dynamic_edge_conv.DynamicEdgeConv(nn: Callable, k: int, aggr: str = 'max', num_workers: int = 1, **kwargs)#

Bases: torch_geometric.nn.conv.MessagePassing

nn#
k#
num_workers#
edge_index = None#
reset_parameters()#
get_edge_index()#
forward(x: torch.Tensor | torch_geometric.typing.PairTensor, batch: torch_geometric.typing.OptTensor | torch_geometric.typing.PairTensor | None = None) → torch.Tensor#
message(x_i: torch.Tensor, x_j: torch.Tensor) → torch.Tensor#
__repr__() → str#
Contents