def __init__(self,
device,
num_nodes,
h_dim,
out_dim,
num_rels,
num_bases=None,
num_hidden_layers=1,
dropout=0,
use_self_loop=False,
low_mem=False,
layer_norm=False):
super(EntityClassify, self).__init__()
self.device = th.device(device if device >= 0 else 'cpu')
self.num_nodes = num_nodes
self.h_dim = h_dim
self.out_dim = out_dim
self.num_rels = num_rels
self.num_bases = None if num_bases < 0 else num_bases
self.num_hidden_layers = num_hidden_layers
self.dropout = dropout
self.use_self_loop = use_self_loop
self.low_mem = low_mem
self.layer_norm = layer_norm
self.layers = nn.ModuleList()
# i2h
self.layers.append(
RelGraphConv(self.h_dim,
self.h_dim,
self.num_rels,
"basis",
self.num_bases,
activation=F.relu,
self_loop=self.use_self_loop,
low_mem=self.low_mem,
dropout=self.dropout))
# h2h
for idx in range(self.num_hidden_layers):
self.layers.append(
RelGraphConv(self.h_dim,
self.h_dim,
self.num_rels,
"basis",
self.num_bases,
activation=F.relu,
self_loop=self.use_self_loop,
low_mem=self.low_mem,
dropout=self.dropout))
# h2o
self.layers.append(
RelGraphConv(self.h_dim,
self.out_dim,
self.num_rels,
"basis",
self.num_bases,
activation=None,
self_loop=self.use_self_loop,
low_mem=self.low_mem))
def __init__(self,
num_nodes,
hidden_dim,
num_rels,
num_layers=2,
regularizer='basis',
num_bases=None,
dropout=0.0):
"""R-GCN连接预测模型
:param num_nodes: int 顶点（实体）数
:param hidden_dim: int 隐含特征维数
:param num_rels: int 关系个数
:param num_layers: int, optional R-GCN层数，默认为2
:param regularizer: str, 'basis'/'bdd' 权重正则化方法，默认为'basis'
:param num_bases: int, optional 基的个数，默认使用关系个数
:param dropout: float, optional Dropout概率，默认为0
"""
super().__init__()
self.embed = nn.Embedding(num_nodes, hidden_dim)
self.layers = nn.ModuleList()
for i in range(num_layers):
self.layers.append(
RelGraphConv(hidden_dim,
hidden_dim,
num_rels,
regularizer,
num_bases,
activation=F.relu if i < num_layers - 1 else None,
self_loop=True,
low_mem=True,
dropout=dropout))
self.score = DistMult(num_rels, hidden_dim)
def __init__(self, in_feat, hid_feat, out_feat, n_hidden_layers, num_rel):
super().__init__()
self.h_embedding = nn.Embedding(in_feat, hid_feat)
self.mlp = MLPPredictor(hid_feat, out_feat)
self.layers = nn.ModuleList([
RelGraphConv(hid_feat,
hid_feat,
num_rel,
low_mem=True,
dropout=0.4) for _ in range(n_hidden_layers)
])
def __init__(self, in_dim, hid_dim, out_dim, n_layers, num_rel, dropout):
super().__init__()
self.in_dim = in_dim
self.hid_dim = hid_dim
self.out_dim = out_dim
self.n_layers = n_layers
self.num_rel = num_rel
self.dropout = dropout
self.h_embedding = nn.Embedding(in_dim, hid_dim)
self.layers = nn.ModuleList([
RelGraphConv(hid_dim,
hid_dim,
num_rel,
regularizer='basis',
num_bases=None,
low_mem=True,
dropout=self.dropout) for _ in range(n_layers)
])
self.bilin_score = BiLinearPredictor(num_rel, hid_dim)