-
Notifications
You must be signed in to change notification settings - Fork 18
/
Copy pathGraphSGAN.py
203 lines (184 loc) · 9.89 KB
/
GraphSGAN.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
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
# -*- coding:utf-8 -*-
from __future__ import print_function
import torch
import numpy as np
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torch.autograd import Variable
from functional import log_sum_exp, pull_away_term
from torch.utils.data import DataLoader,TensorDataset
import sys
from torch.nn.parameter import Parameter
import argparse
from Nets import Generator, Discriminator
import tensorboardX
import os
import random
from FeatureGraphDataset import FeatureGraphDataset
import pickle as pkl
class GraphSGAN(object):
def __init__(self, G, D, dataset, args):
if os.path.exists(args.savedir):
print('Loading model from ' + args.savedir)
self.G = torch.load(os.path.join(args.savedir, 'G.pkl'))
self.D = torch.load(os.path.join(args.savedir, 'D.pkl'))
self.embedding_layer = torch.load(os.path.join(args.savedir, 'embedding.pkl'))
else:
os.makedirs(args.savedir)
self.G = G
self.D = D
self.embedding_layer = nn.Embedding(dataset.n, dataset.d)
self.embedding_layer.weight = Parameter(torch.Tensor(dataset.embbedings))
torch.save(self.G, os.path.join(args.savedir, 'G.pkl'))
torch.save(self.D, os.path.join(args.savedir, 'D.pkl'))
torch.save(self.embedding_layer, os.path.join(args.savedir, 'embedding.pkl'))
self.writer = tensorboardX.SummaryWriter(log_dir=args.logdir)
if args.cuda:
self.G.cuda()
self.D.cuda() # self.embedding_layer is on CPU
self.dataset = dataset
self.Doptim = optim.Adam(self.D.parameters(), lr=args.lr, betas= (args.momentum, 0.999))
self.Goptim = optim.Adam(self.G.parameters(), lr=args.lr, betas = (args.momentum,0.999))
self.args = args
def trainD(self, idf_label, y, idf_unlabel):
x_label, x_unlabel, y = self.make_input(*idf_label), self.make_input(*idf_unlabel), Variable(y, requires_grad = False)
if self.args.cuda:
x_label, x_unlabel, y = x_label.cuda(), x_unlabel.cuda(), y.cuda()
output_label, (mom_un, output_unlabel), output_fake = self.D(x_label, cuda=self.args.cuda), self.D(x_unlabel, cuda=self.args.cuda, feature = True), self.D(self.G(x_unlabel.size()[0], cuda = self.args.cuda).view(x_unlabel.size()).detach(), cuda=self.args.cuda)
logz_label, logz_unlabel, logz_fake = log_sum_exp(output_label), log_sum_exp(output_unlabel), log_sum_exp(output_fake) # log ∑e^x_i
prob_label = torch.gather(output_label, 1, y.unsqueeze(1)) # log e^x_label = x_label
loss_supervised = -torch.mean(prob_label) + torch.mean(logz_label)
loss_unsupervised = 0.5 * (-torch.mean(logz_unlabel) + torch.mean(F.softplus(logz_unlabel)) + # real_data: log Z/(1+Z)
torch.mean(F.softplus(logz_fake)) ) # fake_data: log 1/(1+Z)
entropy = -torch.mean(F.softmax(output_unlabel, dim = 1) * F.log_softmax(output_unlabel, dim = 1))
pt = pull_away_term(mom_un)
loss = loss_supervised + self.args.unlabel_weight * loss_unsupervised + entropy + pt
acc = torch.mean((output_label.max(1)[1] == y).float())
self.Doptim.zero_grad()
loss.backward()
self.Doptim.step()
return loss_supervised.data.cpu().numpy(), loss_unsupervised.data.cpu().numpy(), acc
def trainG(self, idf_unlabel):
x_unlabel = self.make_input(*idf_unlabel)
if self.args.cuda:
x_unlabel = x_unlabel.cuda()
fake = self.G(x_unlabel.size()[0], cuda = self.args.cuda).view(x_unlabel.size())
mom_gen, output_fake = self.D(fake, feature=True, cuda=self.args.cuda)
mom_unlabel, output_unlabel = self.D(x_unlabel, feature=True, cuda=self.args.cuda)
loss_pt = pull_away_term(mom_gen)
mom_gen = torch.mean(mom_gen, dim = 0)
mom_unlabel = torch.mean(mom_unlabel, dim = 0)
loss_fm = torch.mean(torch.abs(mom_gen - mom_unlabel))
loss = loss_fm + loss_pt
self.Goptim.zero_grad()
self.Doptim.zero_grad()
loss.backward()
self.Goptim.step()
return loss.data.cpu().numpy()
def make_input(self, ids, feature, volatile = False):
'''Concatenate feature and embeddings
Args:
feature: Size=>[batch_size, dataset.k], Type=>FloatTensor
ids: Size=>[batch_size], Type=>LongTensor
'''
embedding = self.embedding_layer(Variable(ids, volatile = volatile)).detach() # detach temporarily
return torch.cat((Variable(feature), embedding), dim = 1)
def train(self):
gn = 0
NUM_BATCH = 100
for epoch in range(self.args.epochs):
self.G.train()
self.D.train()
self.D.turn = epoch
loss_supervised = loss_unsupervised = loss_gen = accuracy = 0.
for batch_num in range(NUM_BATCH):
# extract batch from dataset
idf_unlabel1 = self.dataset.unlabel_batch(self.args.batch_size)
idf_unlabel2 = self.dataset.unlabel_batch(self.args.batch_size)
id0, xf, y = self.dataset.label_batch(self.args.batch_size)
# train D
ll, lu, acc = self.trainD((id0, xf), y, idf_unlabel1)
loss_supervised += ll
loss_unsupervised += lu
accuracy += acc
# train G on unlabeled data
lg = self.trainG(idf_unlabel2)
loss_gen += lg
# print and record logs
if (batch_num + 1) % self.args.log_interval == 0:
print('Training: %d / %d' % (batch_num + 1, NUM_BATCH))
gn += 1
self.writer.add_scalars('loss', {'loss_supervised':ll, 'loss_unsupervised':lu, 'loss_gen':lg}, gn)
self.writer.add_histogram('real_feature', self.D(self.make_input(id0, xf, volatile = True).cuda(), cuda=self.args.cuda, feature = True)[0], gn)
self.writer.add_histogram('fake_feature', self.D(self.G(self.args.batch_size, cuda = self.args.cuda), cuda=self.args.cuda, feature = True)[0], gn)
# calculate average loss at the end of an epoch
batch_num += 1
loss_supervised /= batch_num
loss_unsupervised /= batch_num
loss_gen /= batch_num
accuracy /= batch_num
print("Iteration %d, loss_supervised = %.4f, loss_unsupervised = %.4f, loss_gen = %.4f train acc = %.4f" % (epoch, loss_supervised, loss_unsupervised, loss_gen, accuracy))
sys.stdout.flush()
# eval
tmp = self.eval()
print("Eval: correct %d / %d, Acc: %.2f" % (tmp, self.dataset.test_num, tmp * 100. / self.dataset.test_num))
torch.save(self.G, os.path.join(self.args.savedir, 'G.pkl'))
torch.save(self.D, os.path.join(self.args.savedir, 'D.pkl'))
def predict(self, x):
'''predict label in volatile mode
Args:
x: Size=>[batch_size, self.dataset.k + self.dataset.d], Type=>Variable(FloatTensor), volatile
'''
return torch.max(self.D(x, cuda=self.args.cuda), 1)[1].data
def eval(self):
self.G.eval()
self.D.eval()
ids, f, y = self.dataset.test_batch()
x = self.make_input(ids, f, volatile = True)
if self.args.cuda:
x, y = x.cuda(), y.cuda()
pred1 = self.predict(x)
return torch.sum(pred1 == y)
def draw(self, batch_size):
self.G.eval()
return self.G(batch_size, cuda=self.args.cuda)
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='PyTorch GraphS GAN')
parser.add_argument('--batch-size', type=int, default=64, metavar='N',
help='input batch size for training (default: 64)')
parser.add_argument('--epochs', type=int, default=20, metavar='N',
help='number of epochs to train (default: 20)')
parser.add_argument('--lr', type=float, default=0.003, metavar='LR',
help='learning rate (default: 0.003)')
parser.add_argument('--momentum', type=float, default=0.5, metavar='M',
help='SGD momentum (default: 0.5)')
parser.add_argument('--cuda', action='store_true', default=False,
help='CUDA training')
parser.add_argument('--seed', type=int, default=2, metavar='S',
help='random seed (default: 1)')
parser.add_argument('--log-interval', type=int, default=100, metavar='N',
help='how many batches to wait before logging training status')
parser.add_argument('--eval-interval', type=int, default=1, metavar='N',
help='how many batches to wait before evaling training status')
parser.add_argument('--unlabel-weight', type=float, default=0.5, metavar='N',
help='scale factor between labeled and unlabeled data')
parser.add_argument('--logdir', type=str, default='./logfile', metavar='LOG_PATH', help='logfile path, tensorboard format')
parser.add_argument('--savedir', type=str, default='./models', metavar='SAVE_PATH', help = 'saving path, pickle format')
args = parser.parse_args()
args.cuda = args.cuda and torch.cuda.is_available()
np.random.seed(args.seed)
random.seed(args.seed)
torch.manual_seed(args.seed)
if args.cuda:
torch.cuda.manual_seed(args.seed)
# That is how you usually build the dataset
#dataset = CoraDataset(feature_file = './data/cora.features',
# edge_file = './data/cora_edgelist', label_file = './data/cora_label')
#dataset.read_embbedings('./embedding/embedding_line_cora')
#dataset.setting(20, 1000)
# but we load the example of cora
with open('cora.dataset', 'r') as fdata:
dataset = pkl.load(fdata)
gan = GraphSGAN(Generator(200, dataset.k + dataset.d), Discriminator(dataset.k + dataset.d, dataset.m), dataset, args)
gan.train()