def __init__()

in source/sagemaker/sagemaker_graph_fraud_detection/dgl_fraud_detection/model/pytorch.py [0:0]


    def __init__(self,
                 g,
                 in_dim,
                 num_hidden,
                 num_classes,
                 num_layers,
                 heads,
                 activation,
                 feat_drop,
                 attn_drop,
                 alpha,
                 residual):
        super(GAT, self).__init__()
        self.g = g
        self.num_layers = num_layers
        self.gat_layers = nn.ModuleList()
        self.activation = activation
        # input projection (no residual)
        self.gat_layers.append(GATConv(
            in_dim, num_hidden, heads[0],
            feat_drop, attn_drop, alpha, False))
        # hidden layers
        for l in range(1, num_layers):
            # due to multi-head, the in_dim = num_hidden * num_heads
            self.gat_layers.append(GATConv(
                num_hidden * heads[l-1], num_hidden, heads[l],
                feat_drop, attn_drop, alpha, residual))
        # output projection
        self.gat_layers.append(GATConv(
            num_hidden * heads[-2], num_classes, heads[-1],
            feat_drop, attn_drop, alpha, residual))