-
Notifications
You must be signed in to change notification settings - Fork 0
/
model.py
142 lines (114 loc) · 6.09 KB
/
model.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
from layers import StackGraphAttention, InnerProductDecoder
import tensorflow as tf
flags = tf.app.flags
FLAGS = flags.FLAGS
def gaussian_noise_layer(input_layer, std):
"""Noise sampling."""
noise = tf.random_normal(shape=tf.shape(input_layer), mean=0.0, stddev=std, dtype=tf.float32)
return input_layer + noise
def dense(inputs, input_dim, output_dim, name):
"""Used to create a dense layer."""
with tf.variable_scope(name, reuse=None):
tf.set_random_seed(1)
weights = tf.get_variable("weights", shape=[input_dim, output_dim],
initializer=tf.random_normal_initializer(mean=0., stddev=0.01), dtype=tf.float32)
bias = tf.get_variable("bias", shape=[output_dim], initializer=tf.constant_initializer(0.0), dtype=tf.float32)
outputs = tf.add(tf.matmul(inputs, weights), bias, name='matmul')
return outputs
class Model(object):
def __init__(self, **kwargs):
allowed_kwargs = {'name', 'logging'}
for kwarg in kwargs.keys():
assert kwarg in allowed_kwargs, 'Invalid keyword argument: ' + kwarg
for kwarg in kwargs.keys():
assert kwarg in allowed_kwargs, 'Invalid keyword argument: ' + kwarg
name = kwargs.get('name')
if not name:
name = self.__class__.__name__.lower()
self.name = name
logging = kwargs.get('logging', False)
self.logging = logging
self.vars = {}
def _build(self):
raise NotImplementedError
def build(self):
""" Wrapper for _build() """
with tf.variable_scope(self.name):
self._build()
variables = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope=self.name)
self.vars = {var.name: var for var in variables}
class GALG(Model):
def __init__(self, placeholders, feature_length, client_list, **kwargs):
super(GALG, self).__init__(**kwargs)
self.inputs = placeholders['features']
self.input_dim = feature_length
self.adj = placeholders['adj']
self.dropout = placeholders['dropout']
self.node_ids = client_list
self.build()
def _build(self):
with tf.variable_scope('Encoder', reuse=None):
# self.noise = gaussian_noise_layer(self.inputs, 0.1)
self.noise = self.inputs
self.embeddings = StackGraphAttention(input_dim=self.input_dim,
output_dim=FLAGS.hidden2,
adj=self.adj,
node_ids=self.node_ids,
dropout=self.dropout,
n_heads=FLAGS.n_heads,
act=tf.nn.elu,
n_layers=FLAGS.n_layers,
logging=self.logging,
name='e_dense')(self.noise)
self.reconstructions = InnerProductDecoder(input_dim=FLAGS.hidden2,
act=lambda x: x,
logging=self.logging)(self.embeddings)
class VGALG(Model):
def __init__(self, placeholders, feature_length, client_list, **kwargs):
super(VGALG, self).__init__(**kwargs)
self.inputs = placeholders['features']
self.input_dim = feature_length
self.n_samples = len(client_list)
self.adj = placeholders['adj']
self.dropout = placeholders['dropout']
self.node_ids = client_list
self.build()
def _build(self):
with tf.variable_scope('Encoder'):
self.z_mean = StackGraphAttention(input_dim=self.input_dim,
output_dim=FLAGS.hidden2,
adj=self.adj,
node_ids=self.node_ids,
dropout=self.dropout,
n_heads=FLAGS.n_heads,
act=tf.nn.tanh,
n_layers=FLAGS.n_layers,
logging=self.logging,
name='e_dense_1')(self.inputs)
self.z_log_std = StackGraphAttention(input_dim=self.input_dim,
output_dim=FLAGS.hidden2,
adj=self.adj,
node_ids=self.node_ids,
dropout=self.dropout,
n_heads=FLAGS.n_heads,
act=tf.nn.tanh,
n_layers=FLAGS.n_layers,
logging=self.logging,
name='e_dense_2')(self.inputs)
self.embeddings = self.z_mean + tf.random_normal([self.n_samples, FLAGS.hidden2]) * self.z_log_std
self.reconstructions = InnerProductDecoder(input_dim=FLAGS.hidden2,
act=lambda x: x,
logging=self.logging)(self.embeddings)
class Discriminator(Model):
def __init__(self, **kwargs):
super(Discriminator, self).__init__(**kwargs)
self.act = tf.nn.relu
def construct(self, inputs, reuse=False):
with tf.variable_scope('Discriminator'):
if reuse:
tf.get_variable_scope().reuse_variables()
tf.set_random_seed(1)
dc_den1 = tf.nn.relu(dense(inputs, FLAGS.hidden2, FLAGS.hidden3, name='dc_dense_1'))
dc_den2 = tf.nn.relu(dense(dc_den1, FLAGS.hidden3, FLAGS.hidden1, name='dc_dense_2'))
outputs = dense(dc_den2, FLAGS.hidden1, 1, name='dc_outputs')
return outputs