forked from chennnM/GBP
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmultilabel.py
129 lines (106 loc) · 4.1 KB
/
multilabel.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
from __future__ import division
from __future__ import print_function
from utils import load_inductive,mutilabel_f1
import time
import random
import argparse
import numpy as np
import torch
import torch.nn as nn
import torch.optim as optim
import torch.utils.data as Data
from model import GnnBP
import uuid
# Training settings
parser = argparse.ArgumentParser()
parser.add_argument('--seed', type=int, default=20159, help='random seed.')
parser.add_argument('--epochs', type=int, default=1000, help='number of epochs.')
parser.add_argument('--lr', type=float, default=0.01, help='learning rate.')
parser.add_argument('--weight_decay', type=float, default=0, help='weight decay.')
parser.add_argument('--layer', type=int, default=4, help='number of layers.')
parser.add_argument('--hidden', type=int, default=2048, help='hidden dimensions.')
parser.add_argument('--dropout', type=float, default=0.1, help='dropout rate.')
parser.add_argument('--patience', type=int, default=100, help='patience')
parser.add_argument('--data', default='ppi', help='dateset')
parser.add_argument('--dev', type=int, default=0, help='device id')
parser.add_argument('--alpha', type=float, default=0.3, help='decay factor')
parser.add_argument('--rmax', type=float, default=5e-7, help='threshold.')
parser.add_argument('--rrz', type=float, default=0.0, help='rrz.')
parser.add_argument('--bias', default='bn', help='bias.')
parser.add_argument('--batch', type=int, default=1024, help='batch size')
args = parser.parse_args()
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
torch.cuda.manual_seed(args.seed)
print("--------------------------")
print(args)
features_train,features,labels,idx_train,idx_val,idx_test = load_inductive(args.data,args.alpha,args.rmax,args.rrz)
checkpt_file = 'pretrained/'+uuid.uuid4().hex+'.pt'
model = GnnBP(nfeat=features_train.shape[1],
nlayers=args.layer,
nhidden=args.hidden,
nclass=labels.shape[1],
dropout=args.dropout,
bias = args.bias).cuda(args.dev)
optimizer = optim.Adam(model.parameters(), lr=args.lr, weight_decay=args.weight_decay)
loss_fn = nn.BCEWithLogitsLoss()
labels = labels.float()
def evaluate(model,feats_val,labels_val):
model.eval()
with torch.no_grad():
logits = model(feats_val)
f1_mic = mutilabel_f1(labels_val.cpu().numpy(),logits.cpu().numpy())
return f1_mic
torch_dataset = Data.TensorDataset(features_train, labels[idx_train])
loader = Data.DataLoader(dataset=torch_dataset,batch_size=args.batch,shuffle=True,num_workers=40)
def train():
model.train()
loss_list = []
time_epoch = 0
for step, (batch_x, batch_y) in enumerate(loader):
batch_x = batch_x.cuda(args.dev)
batch_y = batch_y.cuda(args.dev)
t1 = time.time()
optimizer.zero_grad()
output = model(batch_x)
loss_train = loss_fn(output, batch_y)
loss_train.backward()
optimizer.step()
time_epoch+=(time.time()-t1)
loss_list.append(loss_train.item())
return np.mean(loss_list),time_epoch
def validate():
return evaluate(model,features[idx_val].cuda(args.dev),labels[idx_val])
def test():
model.load_state_dict(torch.load(checkpt_file))
return evaluate(model,features[idx_test].cuda(args.dev),labels[idx_test])
train_time = 0
bad_counter = 0
best = 0
best_epoch = 0
for epoch in range(args.epochs):
loss_tra,train_ep = train()
f1_val = validate()
train_time+=train_ep
if(epoch+1)%100 == 0:
print('Epoch:{:04d}'.format(epoch+1),
'train',
'loss:{:.3f}'.format(loss_tra),
'| val',
'acc:{:.3f}'.format(f1_val),
'| cost{:.3f}'.format(train_time))
if f1_val > best:
best = f1_val
best_epoch = epoch
torch.save(model.state_dict(), checkpt_file)
bad_counter = 0
else:
bad_counter += 1
if bad_counter == args.patience:
break
f1_test = test()
print("Train cost: {:.4f}s".format(train_time))
print('Load {}th epoch'.format(best_epoch))
print("Test f1:{:.3f}".format(f1_test))
print("--------------------------")