Source code for torch_geometric.nn.conv.sage_conv

from typing import Optional, Tuple, Union

import torch
import torch.nn.functional as F
from torch import Tensor
from torch.nn import LSTM
from torch_scatter import scatter
from torch_sparse import SparseTensor, matmul

from torch_geometric.nn.conv import MessagePassing
from torch_geometric.nn.dense.linear import Linear
from torch_geometric.typing import Adj, OptPairTensor, Size
from torch_geometric.utils import to_dense_batch

[docs]class SAGEConv(MessagePassing): r"""The GraphSAGE operator from the `"Inductive Representation Learning on Large Graphs" <>`_ paper .. math:: \mathbf{x}^{\prime}_i = \mathbf{W}_1 \mathbf{x}_i + \mathbf{W}_2 \cdot \mathrm{mean}_{j \in \mathcal{N(i)}} \mathbf{x}_j If :obj:`project = True`, then :math:`\mathbf{x}_j` will first get projected via .. math:: \mathbf{x}_j \leftarrow \sigma ( \mathbf{W}_3 \mathbf{x}_j + \mathbf{b}) as described in Eq. (3) of the paper. Args: in_channels (int or tuple): Size of each input sample, or :obj:`-1` to derive the size from the first input(s) to the forward method. A tuple corresponds to the sizes of source and target dimensionalities. out_channels (int): Size of each output sample. aggr (string, optional): The aggregation scheme to use (:obj:`"mean"`, :obj:`"max"`, :obj:`"lstm"`). (default: :obj:`"add"`) normalize (bool, optional): If set to :obj:`True`, output features will be :math:`\ell_2`-normalized, *i.e.*, :math:`\frac{\mathbf{x}^{\prime}_i} {\| \mathbf{x}^{\prime}_i \|_2}`. (default: :obj:`False`) root_weight (bool, optional): If set to :obj:`False`, the layer will not add transformed root node features to the output. (default: :obj:`True`) project (bool, optional): If set to :obj:`True`, the layer will apply a linear transformation followed by an activation function before aggregation (as described in Eq. (3) of the paper). (default: :obj:`False`) bias (bool, optional): If set to :obj:`False`, the layer will not learn an additive bias. (default: :obj:`True`) **kwargs (optional): Additional arguments of :class:`torch_geometric.nn.conv.MessagePassing`. Shapes: - **inputs:** node features :math:`(|\mathcal{V}|, F_{in})` or :math:`((|\mathcal{V_s}|, F_{s}), (|\mathcal{V_t}|, F_{t}))` if bipartite, edge indices :math:`(2, |\mathcal{E}|)` - **outputs:** node features :math:`(|\mathcal{V}|, F_{out})` or :math:`(|\mathcal{V_t}|, F_{out})` if bipartite """ def __init__( self, in_channels: Union[int, Tuple[int, int]], out_channels: int, aggr: str = 'mean', normalize: bool = False, root_weight: bool = True, project: bool = False, bias: bool = True, **kwargs, ): kwargs['aggr'] = aggr if aggr != 'lstm' else None super().__init__(**kwargs) self.in_channels = in_channels self.out_channels = out_channels self.normalize = normalize self.root_weight = root_weight self.project = project if isinstance(in_channels, int): in_channels = (in_channels, in_channels) if self.project: self.lin = Linear(in_channels[0], in_channels[0], bias=True) if self.aggr is None: self.fuse = False # No "fused" message_and_aggregate. self.lstm = LSTM(in_channels[0], in_channels[0], batch_first=True) self.lin_l = Linear(in_channels[0], out_channels, bias=bias) if self.root_weight: self.lin_r = Linear(in_channels[1], out_channels, bias=False) self.reset_parameters()
[docs] def reset_parameters(self): if self.project: self.lin.reset_parameters() if self.aggr is None: self.lstm.reset_parameters() self.lin_l.reset_parameters() if self.root_weight: self.lin_r.reset_parameters()
[docs] def forward(self, x: Union[Tensor, OptPairTensor], edge_index: Adj, size: Size = None) -> Tensor: """""" if isinstance(x, Tensor): x: OptPairTensor = (x, x) if self.project and hasattr(self, 'lin'): x = (self.lin(x[0]).relu(), x[1]) # propagate_type: (x: OptPairTensor) out = self.propagate(edge_index, x=x, size=size) out = self.lin_l(out) x_r = x[1] if self.root_weight and x_r is not None: out += self.lin_r(x_r) if self.normalize: out = F.normalize(out, p=2., dim=-1) return out
def message(self, x_j: Tensor) -> Tensor: return x_j def message_and_aggregate(self, adj_t: SparseTensor, x: OptPairTensor) -> Tensor: adj_t = adj_t.set_value(None, layout=None) return matmul(adj_t, x[0], reduce=self.aggr) def aggregate(self, x: Tensor, index: Tensor, ptr: Optional[Tensor] = None, dim_size: Optional[int] = None) -> Tensor: if self.aggr is not None: return scatter(x, index, dim=self.node_dim, dim_size=dim_size, reduce=self.aggr) # LSTM aggregation: if ptr is None and not torch.all(index[:-1] <= index[1:]): raise ValueError(f"Can not utilize LSTM-style aggregation inside " f"'{self.__class__.__name__}' in case the " f"'edge_index' tensor is not sorted by columns. " f"Run 'sort_edge_index(..., sort_by_row=False)' " f"in a pre-processing step.") x, mask = to_dense_batch(x, batch=index, batch_size=dim_size) out, _ = self.lstm(x) return out[:, -1] def __repr__(self) -> str: aggr = self.aggr if self.aggr is not None else 'lstm' return (f'{self.__class__.__name__}({self.in_channels}, ' f'{self.out_channels}, aggr={aggr})')