def __init__(self,
input_dim,
output_dim,
model_size="small",
neigh_input_dim=None,
dropout=0.,
bias=False,
act=tf.nn.relu,
name=None,
concat=False,
**kwargs):
super(MeanPoolingAggregator, self).__init__(**kwargs)
self.dropout = dropout
self.bias = bias
self.act = act
self.concat = concat
if neigh_input_dim is None:
neigh_input_dim = input_dim
if name is not None:
name = '/' + name
else:
name = ''
if model_size == "small":
hidden_dim = self.hidden_dim = 512
elif model_size == "big":
hidden_dim = self.hidden_dim = 1024
self.mlp_layers = []
self.mlp_layers.append(
Dense(input_dim=neigh_input_dim,
output_dim=hidden_dim,
act=tf.nn.relu,
dropout=dropout,
sparse_inputs=False,
logging=self.logging))
with tf.variable_scope(self.name + name + '_vars'):
self.vars['neigh_weights'] = glorot([hidden_dim, output_dim],
name='neigh_weights')
self.vars['self_weights'] = glorot([input_dim, output_dim],
name='self_weights')
if self.bias:
self.vars['bias'] = zeros([self.output_dim], name='bias')
if self.logging:
self._log_vars()
self.input_dim = input_dim
self.output_dim = output_dim
self.neigh_input_dim = neigh_input_dim
def __init__(self,
input_dim,
output_dim,
model_size="small",
neigh_input_dim=None,
dropout=0.,
bias=False,
act=tf.nn.relu,
name=None,
concat=False,
**kwargs):
super(SeqAggregator, self).__init__(**kwargs)
self.dropout = dropout
self.bias = bias
self.act = act
self.concat = concat
if neigh_input_dim is None:
neigh_input_dim = input_dim
if name is not None:
name = '/' + name
else:
name = ''
if model_size == "small":
hidden_dim = self.hidden_dim = 128
elif model_size == "big":
hidden_dim = self.hidden_dim = 256
with tf.variable_scope(self.name + name + '_vars'):
self.vars['neigh_weights'] = glorot([hidden_dim, output_dim],
name='neigh_weights')
self.vars['self_weights'] = glorot([input_dim, output_dim],
name='self_weights')
if self.bias:
self.vars['bias'] = zeros([self.output_dim], name='bias')
if self.logging:
self._log_vars()
self.input_dim = input_dim
self.output_dim = output_dim
self.neigh_input_dim = neigh_input_dim
self.cell = tf.contrib.rnn.BasicLSTMCell(self.hidden_dim)
def __init__(self,
input_dim,
output_dim,
neigh_input_dim=None,
dropout=0.,
bias=False,
act=tf.nn.relu,
name=None,
concat=False,
**kwargs):
super(MeanAggregator, self).__init__(**kwargs)
# print('========== init MeanAggregator =========')
# import traceback
# traceback.print_stack()
self.dropout = dropout
self.bias = bias
self.act = act
self.concat = concat
if neigh_input_dim is None:
neigh_input_dim = input_dim
if name is not None:
name = '/' + name
else:
name = ''
with tf.variable_scope(self.name + name + '_vars'):
self.vars['neigh_weights'] = glorot([neigh_input_dim, output_dim],
name='neigh_weights')
self.vars['self_weights'] = glorot([input_dim, output_dim],
name='self_weights')
# self.vars['alpha'] = glorot([9], name='alpha') # 增加，居然是硬编码9！
if self.bias:
self.vars['bias'] = zeros([self.output_dim], name='bias')
if self.logging:
self._log_vars()
self.input_dim = input_dim
self.output_dim = output_dim
def config_embeds(self, adj_info, features, embed_dim):
zero_padding = tf.zeros([1, embed_dim], name="zero_padding")
embeds = glorot([len(adj_info), embed_dim], name="node_emebddings")
self.embeds = tf.concat([zero_padding, embeds], axis=0)
if features is None:
self.features = self.embeds
else:
self.features = tf.Variable(tf.constant(features,
dtype=tf.float32),
trainable=False)
self.features = tf.concat([self.embeds, self.features], axis=1)
def __init__(self,
input_dim,
output_dim,
neigh_input_dim=None,
dropout=0.,
bias=False,
act=tf.nn.relu,
name=None,
concat=False,
**kwargs):
super(CrossAggregator, self).__init__(**kwargs)
self.dropout = dropout
self.bias = bias
self.act = act
self.concat = concat
if neigh_input_dim is None:
neigh_input_dim = input_dim
if name is not None:
name = '/' + name
else:
name = ''
# 写死了
atten_dim = 64
num_attr = 9 # 嘉琪说10 column初始为零，9的话，gather_nd 会给零值
with tf.variable_scope(self.name + name + '_vars'):
self.vars['neigh_weights'] = glorot([neigh_input_dim, output_dim],
name='neigh_weights')
self.vars['self_weights'] = glorot([input_dim, output_dim],
name='self_weights')
# self.vars['initial_weights'] = glorot([102, output_dim],
# name='initial_weights')
self.vars['alpha'] = glorot([num_attr], name='alpha')
self.vars['self_atten'] = glorot([input_dim, atten_dim],
name='self_atten')
self.vars['neigh_atten'] = glorot([neigh_input_dim, atten_dim],
name='neigh_atten')
self.vars['v'] = glorot([atten_dim, 1], name='v')
if self.bias:
self.vars['bias'] = zeros([self.output_dim], name='bias')
if self.logging:
self._log_vars()
self.input_dim = input_dim
self.output_dim = output_dim
def __init__(self,
input_dim,
output_dim,
neigh_input_dim=None,
dropout=0.,
bias=False,
act=tf.nn.relu,
name=None,
concat=False,
**kwargs):
super(GCNAggregator, self).__init__(**kwargs)
self.dropout = dropout
self.bias = bias
self.act = act
self.concat = concat
if neigh_input_dim is None:
neigh_input_dim = input_dim
if name is not None:
name = '/' + name
else:
name = ''
with tf.variable_scope(self.name + name + '_vars'):
self.vars['weights'] = glorot([neigh_input_dim, output_dim],
name='neigh_weights')
# self.vars['alpha'] = glorot([9], name='alpha') # 增加
if self.bias:
self.vars['bias'] = zeros([self.output_dim], name='bias')
if self.logging:
self._log_vars()
self.input_dim = input_dim
self.output_dim = output_dim
def __init__(self,
input_dim,
output_dim,
neigh_input_dim=None,
dropout=0,
bias=False,
act=tf.nn.tanh,
name=None,
concat=False,
**kwargs):
super(GeniePathAggregator, self).__init__(**kwargs)
self.dropout = dropout
self.bias = bias
self.act = act
self.concat = concat
if neigh_input_dim is None:
neigh_input_dim = input_dim
if name is not None:
name = '/' + name
else:
name = ''
with tf.variable_scope(self.name + name + '_vars'):
self.vars['W_s'] = glorot([input_dim, output_dim], name='W_s')
self.vars['W_d'] = glorot([neigh_input_dim, output_dim],
name='W_d')
self.vars['v'] = glorot([output_dim, 1], name='v')
self.vars['W_t'] = glorot([input_dim, output_dim], name='W_t')
self.vars['b_t'] = glorot([1, output_dim], name='b_t')
self.vars['W_i'] = glorot([input_dim, output_dim], name='W_i')
self.vars['b_i'] = glorot([1, output_dim], name='b_i')
self.vars['W_f'] = glorot([input_dim, output_dim], name='W_f')
self.vars['b_f'] = glorot([1, output_dim], name='b_f')
self.vars['W_o'] = glorot([input_dim, output_dim], name='W_o')
self.vars['b_o'] = glorot([1, output_dim], name='b_o')
self.vars['W_c'] = glorot([input_dim, output_dim], name='W_c')
self.vars['b_c'] = glorot([1, output_dim], name='b_c')
if self.logging:
self._log_vars()
self.input_dim = input_dim
self.output_dim = output_dim