异构图GCN实现代码

下面是异构图GCN的PyTorch实现代码，其中包括了节点嵌入、异构邻居采样、GCN层等部分：

import torch
import torch.nn as nn
import torch.nn.functional as F
from dgl.nn.pytorch.conv import SAGEConv

class HeteroGCN(nn.Module):
    def __init__(self, in_feats, hidden_size, out_feats, num_layers, hetero_graph, agg_mode='mean'):
        super(HeteroGCN, self).__init__()
        self.in_feats = in_feats
        self.hidden_size = hidden_size
        self.out_feats = out_feats
        self.num_layers = num_layers
        self.hetero_graph = hetero_graph
        self.agg_mode = agg_mode
        
        # 节点嵌入层
        self.embed_layers = nn.ModuleDict({
            ntype: nn.Linear(in_feats, hidden_size) for ntype in hetero_graph.ntypes
        })
        
        # GCN层
        self.layers = nn.ModuleList([
            SAGEConv(hidden_size, hidden_size, agg_mode)
            for _ in range(num_layers - 1)
        ])
        self.layers.append(SAGEConv(hidden_size, out_feats, agg_mode))
        
    def forward(self, blocks, x):
        h = {ntype: self.embed_layers[ntype](x[ntype]) for ntype in self.hetero_graph.ntypes}
        for i, layer in enumerate(self.layers):
            # 异构邻居采样
            h_dst = {etype: h[ntype] for ntype, etype, _ in self.hetero_graph.canonical_etypes}
            h = layer(blocks[i], h_dst)
            if i != self.num_layers - 1:
                h = {ntype: F.relu(h[ntype]) for ntype in self.hetero_graph.ntypes}
        return h

其中，`in_feats`表示输入特征的维度，`hidden_size`表示GCN层的隐藏层大小，`out_feats`表示输出特征的维度，`num_layers`表示GCN层数。`hetero_graph`是一个DGL异构图对象，`agg_mode`表示邻居聚合方式。在`__init__`方法中，我们首先定义了节点嵌入层和GCN层，然后在`forward`方法中进行了异构邻居采样和GCN的前向传播操作。