-
Notifications
You must be signed in to change notification settings - Fork 4
/
main_my_variant3.py
365 lines (301 loc) · 17.2 KB
/
main_my_variant3.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
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
from __future__ import print_function
import matplotlib
matplotlib.use('Agg')
import argparse
import torch
import torch.utils.data
from torch import nn, optim
from torch.autograd import Variable
import time
import random
import os
import math
import pickle
import scipy.sparse as sp
import matplotlib.pyplot as plt
import matplotlib.ticker as ticker
import numpy as np
from Dataset_CDAE import Dataset
from model_my_variant3 import CDAE, Discriminator, save_embedding_process
import time
import itertools
import pandas as pd
from my_utils import *
from scipy.sparse import csr_matrix
'''
variant of my. Set W1, W2 are two-layer non-linear transformation
'''
method_name = 'my_variant3'
# os.environ['CUDA_VISIBLE_DEVICES'] = '1'
topK_list = [5, 10]
parser = argparse.ArgumentParser()
parser.add_argument('--prior', type=str, default='Gaussian', help='Gaussian, MVGaussian')
parser.add_argument('--seed', type=int, default=72, help='Random seed.')
parser.add_argument('--epochs', type=int, default=300, help='Number of epochs to train.')
parser.add_argument('--batch', type=int, default=256, help='batch size.')
parser.add_argument('--emb_size', type=int, default=200, help='embed size.')
parser.add_argument('--lr', type=float, default=0.001, help='Initial learning rate.')
parser.add_argument('--weight_decay', type=float, default=0.001, help='Weight decay (L2 loss on parameters).')
parser.add_argument('--dropout', type=float, default=0.5, help='Dropout rate (1 - keep probability).')
parser.add_argument('--log', type=str, default='logs/{}'.format(method_name), help='log directory')
parser.add_argument('--pos-weight', type=float, default=1.0, help='weight for positive samples')
parser.add_argument('--reg', type=float, default=1.0, help='lambda reg always 0.0 for this variant')
parser.add_argument('--self', type=float, default=1.0, help='lambda rec')
parser.add_argument('--cross', type=float, default=1.0, help='lambda rec')
parser.add_argument('--gan', type=float, default=1.0, help='lambda gan')
parser.add_argument('--d_epoch', type=int, default=2, help='d epoch')
parser.add_argument('--t_percent', type=float, default=1.0, help='target percent')
parser.add_argument('--s_percent', type=float, default=1.0, help='source percent')
parser.add_argument('--dataset', type=str, default='amazon3', help='amazon')
args = parser.parse_args()
args.cuda = torch.cuda.is_available()
kwargs = {'num_workers': 1, 'pin_memory': True} if args.cuda else {}
def sigmoid(x):
return 1 / (1 + np.exp(-x))
def main():
log = os.path.join(args.log, '{}_{}_{}_{}_{}_{}_{}_{}_{}'.format(args.dataset, args.emb_size, args.weight_decay, args.self,
args.cross, args.gan, args.reg, args.t_percent, args.s_percent))
if os.path.isdir(log):
print("%s already exist. are you sure to override? Ok, I'll wait for 5 seconds. Ctrl-C to abort." % log)
time.sleep(5)
os.system('rm -rf %s/' % log)
os.makedirs(log)
print("made the log directory", log)
print('preparing data...')
dataset = Dataset(args.batch, dataset=args.dataset)
NUM_USER = dataset.num_user
NUM_MOVIE = dataset.num_movie
NUM_BOOK = dataset.num_book
print('Preparing the training data......')
# prepare data for X
row, col = dataset.get_part_train_indices('movie', args.s_percent)
values = np.ones(row.shape[0])
user_x = csr_matrix((values, (row, col)), shape=(NUM_USER, NUM_MOVIE)).toarray()
# prepare data fot Y
row, col = dataset.get_part_train_indices('book', args.t_percent)
values = np.ones(row.shape[0])
user_y = csr_matrix((values, (row, col)), shape=(NUM_USER, NUM_BOOK)).toarray()
print('Preparing the training data over......')
# for shared user one-hot representation
user_id = np.arange(NUM_USER).reshape([NUM_USER, 1])
user_x = torch.FloatTensor(user_x)
user_y = torch.FloatTensor(user_y)
train_loader = torch.utils.data.DataLoader(torch.from_numpy(user_id),
batch_size=args.batch,
shuffle=True, **kwargs)
save_loader = torch.utils.data.DataLoader(torch.from_numpy(user_id),
batch_size=args.batch,
shuffle=False, **kwargs)
pos_weight = torch.FloatTensor([args.pos_weight])
if args.cuda:
pos_weight = pos_weight.cuda()
model = CDAE(NUM_USER=NUM_USER, NUM_MOVIE=NUM_MOVIE, NUM_BOOK=NUM_BOOK,
EMBED_SIZE=args.emb_size, dropout=args.dropout)
disc1 = Discriminator(args.emb_size, args.dropout)
disc2 = Discriminator(args.emb_size, args.dropout)
optimizer_g = optim.Adam(model.parameters(), lr=args.lr, weight_decay=args.weight_decay)
optimizer_d = optim.SGD(itertools.chain(disc1.parameters(), disc2.parameters()), lr=args.lr, weight_decay=args.weight_decay)
BCEWL = torch.nn.BCEWithLogitsLoss(reduction='none', pos_weight=pos_weight)
BCEL = torch.nn.BCEWithLogitsLoss(reduction='none')
if args.cuda:
model = model.cuda()
disc1 = disc1.cuda()
disc2 = disc2.cuda()
# prepare data fot test process
movie_vali, movie_test, movie_nega = dataset.movie_vali, dataset.movie_test, dataset.movie_nega
book_vali, book_test, book_nega = dataset.book_vali, dataset.book_test, dataset.book_nega
feed_data = {}
feed_data['fts1'] = user_x
feed_data['fts2'] = user_y
feed_data['movie_vali'] = movie_vali
feed_data['book_vali'] = book_vali
feed_data['movie_test'] = movie_test
feed_data['book_test'] = book_test
feed_data['movie_nega'] = movie_nega
feed_data['book_nega'] = book_nega
best_hr1, best_ndcg1, best_mrr1 = 0.0, 0.0, 0.0
best_hr2, best_ndcg2, best_mrr2 = 0.0, 0.0, 0.0
val_hr1_list, val_ndcg1_list, val_mrr1_list = [], [], []
val_hr2_list, val_ndcg2_list, val_mrr2_list = [], [], []
G_loss_list = []
D_loss_list = []
reg_loss_list = []
loss_list = []
self_loss_list = []
cross_loss_list = []
JRL_loss_list = []
for epoch in range(args.epochs):
model.train()
batch_G_loss_list = []
batch_D_loss_list = []
batch_loss_list = []
batch_reg_loss_list = []
batch_self_loss_list = []
batch_cross_loss_list = []
batch_JRL_loss_list = []
for batch_idx, data in enumerate(train_loader):
data = data.reshape([-1])
if (batch_idx+1) % (args.d_epoch + 1) == 0:
optimizer_d.zero_grad()
prior = torch.from_numpy(np.random.normal(0, 1.0, size=[data.shape[0], args.emb_size])).float()
if args.cuda:
prior = prior.cuda()
if args.cuda:
batch_user = data.cuda()
batch_user_x = user_x[data].cuda()
batch_user_y = user_y[data].cuda()
else:
batch_user = data
batch_user_x = user_x[data]
batch_user_y = user_y[data]
_, _, _, _, z_x, z_y, z_x_reg_loss, z_y_reg_loss = model.forward(batch_user, batch_user_x, batch_user_y)
true1 = disc1(prior).reshape([-1])
true2 = disc2(prior).reshape([-1])
fake1 = disc1(z_x).reshape([-1])
fake2 = disc2(z_y).reshape([-1])
dis_loss1 = BCEL(torch.cat([true1, fake1], 0),
torch.cat([torch.ones_like(true1), torch.zeros_like(fake1)], 0)).sum()
dis_loss2 = BCEL(torch.cat([true2, fake2], 0),
torch.cat([torch.ones_like(true2), torch.zeros_like(fake2)], 0)).sum()
D_loss = dis_loss1 + dis_loss2
D_loss.backward()
optimizer_d.step()
batch_D_loss_list.append(D_loss.item())
else:
optimizer_g.zero_grad()
if args.cuda:
batch_user = data.cuda()
batch_user_x = user_x[data].cuda()
batch_user_y = user_y[data].cuda()
else:
batch_user_x = user_x[data]
batch_user_y = user_y[data]
pred_x, pred_y, pred_x2y, pred_y2x, z_x, z_y, z_x_reg_loss, z_y_reg_loss = model.forward(batch_user, batch_user_x, batch_user_y)
loss_x = BCEWL(pred_x, batch_user_x).sum()
loss_y = BCEWL(pred_y, batch_user_y).sum()
loss_x2y = BCEWL(pred_x2y, batch_user_y).sum()
loss_y2x = BCEWL(pred_y2x, batch_user_x).sum()
reg_loss = z_x_reg_loss.sum() + z_y_reg_loss.sum()
fake_x = disc1(z_x).reshape([-1])
fake_y = disc2(z_y).reshape([-1])
G_loss1 = BCEL(fake_x, torch.ones_like(fake_x)).sum()
G_loss2 = BCEL(fake_y, torch.ones_like(fake_y)).sum()
G_loss = G_loss1+G_loss2
# get plot JRL loss
JRL_loss = args.self * (loss_x + loss_y) + args.cross * (loss_x2y + loss_y2x) + args.reg*reg_loss
batch_JRL_loss_list.append(JRL_loss.item()/args.batch)
# get the whole loss
loss = G_loss + args.self * (loss_x + loss_y) + args.cross * (loss_x2y + loss_y2x) + args.reg*reg_loss
loss.backward()
optimizer_g.step()
batch_G_loss_list.append(G_loss.item())
batch_self_loss_list.append(args.self * (loss_x + loss_y).item())
batch_cross_loss_list.append(args.cross * (loss_x2y + loss_y2x).item())
batch_loss_list.append(loss.item())
batch_reg_loss_list.append(args.reg*reg_loss.item())
epoch_G_loss = np.mean(batch_G_loss_list)
epoch_D_loss = np.mean(batch_D_loss_list)
epoch_reg_loss = np.mean(batch_reg_loss_list)
epoch_loss = np.mean(batch_loss_list)
epoch_self_loss = np.mean(batch_self_loss_list)
epoch_cross_loss = np.mean(batch_cross_loss_list)
epoch_JRL_Loss = np.mean(batch_JRL_loss_list)
G_loss_list.append(epoch_G_loss)
D_loss_list.append(epoch_D_loss)
loss_list.append(epoch_loss)
self_loss_list.append(epoch_self_loss)
cross_loss_list.append(epoch_cross_loss)
reg_loss_list.append(epoch_reg_loss)
JRL_loss_list.append(epoch_JRL_Loss)
print('epoch:{}, self loss:{:.4f}, cross loss:{:.4f}, G loss:{:.4f}, D loss:{:.4f}, reg loss:{:.4f}'.format(epoch,epoch_self_loss,
epoch_cross_loss,
epoch_G_loss,
epoch_D_loss,
epoch_reg_loss))
with open(log + '/tmp.txt', 'a') as f:
f.write('epoch:{}, self loss:{:.4f}, cross loss:{:.4f}, G loss:{:.4f}, D loss:{:.4f}, reg loss:{:.4f}'.format(
epoch,
epoch_self_loss, epoch_cross_loss, epoch_G_loss, epoch_D_loss, epoch_reg_loss))
if epoch % 1 == 0:
model.eval()
avg_hr1, avg_ndcg1, avg_mrr1, avg_hr2, avg_ndcg2, avg_mrr2 = test_process(model, train_loader, feed_data,
args.cuda, topK_list[1], mode='val')
val_hr1_list.append(avg_hr1)
val_ndcg1_list.append(avg_ndcg1)
val_mrr1_list.append(avg_mrr1)
val_hr2_list.append(avg_hr2)
val_ndcg2_list.append(avg_ndcg2)
val_mrr2_list.append(avg_mrr2)
print('test: movie: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}, book: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}'
.format(avg_hr1, avg_ndcg1, avg_mrr1, avg_hr2, avg_ndcg2, avg_mrr2))
with open(log + '/tmp.txt', 'a') as f:
f.write('test: movie: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}, book: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}'
.format(avg_hr1, avg_ndcg1, avg_mrr1, avg_hr2, avg_ndcg2, avg_mrr2))
if avg_hr1 > best_hr1:
best_hr1 = avg_hr1
torch.save(model.state_dict(), os.path.join(log, 'best_hr1.pkl'))
if avg_ndcg1 > best_ndcg1:
torch.save(model.state_dict(), os.path.join(log, 'best_ndcg1.pkl'))
best_ndcg1 = avg_ndcg1
if avg_mrr1 > best_mrr1:
torch.save(model.state_dict(), os.path.join(log, 'best_mrr1.pkl'))
best_mrr1 = avg_mrr1
if avg_hr2 > best_hr2:
torch.save(model.state_dict(), os.path.join(log, 'best_hr2.pkl'))
best_hr2 = avg_hr2
if avg_ndcg2 > best_ndcg2:
torch.save(model.state_dict(), os.path.join(log, 'best_ndcg2.pkl'))
best_ndcg2 = avg_ndcg2
if avg_mrr2 > best_mrr2:
torch.save(model.state_dict(), os.path.join(log, 'best_mrr2.pkl'))
best_mrr2 = avg_mrr2
print('best val movie: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}, book: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}'
.format(best_hr1, best_ndcg1, best_mrr1, best_hr2, best_ndcg2, best_mrr2))
with open(log + '/tmp.txt', 'a') as f:
f.write('best val movie: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}, book: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}'
.format(best_hr1, best_ndcg1, best_mrr1, best_hr2, best_ndcg2, best_mrr2))
# # save necessary plot loss and validation metric
# pickle.dump(JRL_loss_list, open(os.path.join(os.getcwd(), 'saved_embeddings', method_name,
# '{}_{}_JRL_loss_list'.format(args.dataset, args.gan)), 'wb'))
# pickle.dump(val_hr1_list, open(os.path.join(os.getcwd(), 'saved_embeddings', method_name,
# '{}_{}_val_hr1_list'.format(args.dataset, args.gan)), 'wb'))
# pickle.dump(val_ndcg1_list, open(os.path.join(os.getcwd(), 'saved_embeddings', method_name,
# '{}_{}_val_ndcg1_list'.format(args.dataset, args.gan)), 'wb'))
# pickle.dump(val_mrr1_list, open(os.path.join(os.getcwd(), 'saved_embeddings', method_name,
# '{}_{}_val_mrr1_list'.format(args.dataset, args.gan)), 'wb'))
# pickle.dump(val_hr2_list, open(os.path.join(os.getcwd(), 'saved_embeddings', method_name,
# '{}_{}_val_hr2_list'.format(args.dataset, args.gan)), 'wb'))
# pickle.dump(val_ndcg2_list, open(os.path.join(os.getcwd(), 'saved_embeddings', method_name,
# '{}_{}_val_ndcg2_list'.format(args.dataset, args.gan)), 'wb'))
# pickle.dump(val_mrr2_list, open(os.path.join(os.getcwd(), 'saved_embeddings', method_name,
# '{}_{}_val_mrr2_list'.format(args.dataset, args.gan)), 'wb'))
print('Val process over!')
print('Test process......')
for topK in topK_list:
model.load_state_dict(torch.load(os.path.join(log, 'best_hr1.pkl')))
test_hr1, _, _, _, _, _ = test_process(model, train_loader, feed_data, args.cuda, topK, mode='test')
# if topK==10:
# user_embedding1, user_embedding2 = save_embedding_process(model, save_loader, feed_data, args.cuda)
# pickle.dump(user_embedding1, open(
# os.path.join(os.getcwd(), 'saved_embeddings', method_name, '{}_Z_movie.pkl'.format(args.dataset)), 'wb'))
# pickle.dump(user_embedding2, open(
# os.path.join(os.getcwd(), 'saved_embeddings', method_name, '{}_Z_book.pkl'.format(args.dataset)), 'wb'))
model.load_state_dict(torch.load(os.path.join(log, 'best_ndcg1.pkl')))
_, test_ndcg1, _, _, _, _ = test_process(model, train_loader, feed_data, args.cuda, topK, mode='test')
model.load_state_dict(torch.load(os.path.join(log, 'best_mrr1.pkl')))
_, _, test_mrr1, _, _, _ = test_process(model, train_loader, feed_data, args.cuda, topK, mode='test')
model.load_state_dict(torch.load(os.path.join(log, 'best_hr2.pkl')))
_, _, _, test_hr2, _, _ = test_process(model, train_loader, feed_data, args.cuda, topK, mode='test')
model.load_state_dict(torch.load(os.path.join(log, 'best_ndcg2.pkl')))
_, _, _, _, test_ndcg2, _ = test_process(model, train_loader, feed_data, args.cuda, topK, mode='test')
model.load_state_dict(torch.load(os.path.join(log, 'best_mrr2.pkl')))
_, _, _, _, _, test_mrr2 = test_process(model, train_loader, feed_data, args.cuda, topK, mode='test')
print('Test TopK:{} ---> movie: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}, book: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}'
.format(topK, test_hr1, test_ndcg1, test_mrr1, test_hr2, test_ndcg2, test_mrr2))
with open(log + '/tmp.txt', 'a') as f:
f.write('Test TopK:{} ---> movie: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}, book: hr:{:.4f},ndcg:{:.4f},mrr:{:.4f}'
.format(topK, test_hr1, test_ndcg1, test_mrr1, test_hr2, test_ndcg2, test_mrr2))
if __name__ == "__main__":
print(args)
main()
print(args)