-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtrain_1.py
86 lines (79 loc) · 3.72 KB
/
train_1.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
from torch import optim
from torch.utils.tensorboard import SummaryWriter
from model_1 import *
from tqdm import tqdm
if __name__ == '__main__':
batch_size = 512
epoch = 300
learning_rate = 1e-3
sequence_length = 64 # 截取的序列长度
cov_size = 256 # 卷积核数
embedding_size = 50 # 词向量化维度
region_size = 3 # 卷积核高度(如果muti_conv==true, 则卷积核高度为(region_size,region_size+1,region_size+2))
min_word_frequency = 1 # 最小词频
participle = False # 是否分词
muti_conv = False # 是否采用多卷积
keywords = True # 是否使用keywords
record = False # 是否保存
'''
b=4_lr=1e-3_sl=20_cs=10_es=100_rs=5_m=1
'''
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print(device)
# 读取并预处理数据
train_inputs, train_labels, vocab_size = read_data(sequence_length=sequence_length, drop_num=min_word_frequency, participle=participle, key_words=keywords)
test_inputs, test_labels, _ = read_data(train=False,sequence_length=sequence_length, drop_num=min_word_frequency, participle=participle, key_words=keywords)
train_len = len(train_labels)
test_len = len(test_labels)
labels = train_labels + test_labels
labels, label2idx = label_digitize(labels)
num_classes = len(label2idx)
train_labels = labels[:len(train_labels)]
test_labels = labels[len(train_labels):]
train_inputs = torch.LongTensor(train_inputs).to(device)
train_labels = torch.LongTensor(train_labels).to(device)
test_inputs = torch.LongTensor(test_inputs).to(device)
test_labels = torch.LongTensor(test_labels).to(device)
train_dataset = Data.TensorDataset(train_inputs, train_labels)
test_dataset = Data.TensorDataset(test_inputs, test_labels)
train_data_loader = Data.DataLoader(train_dataset, batch_size, True)
test_data_loader = Data.DataLoader(test_dataset, batch_size, True)
# 训练
textcnn = TextCNN(num_classes, vocab_size, cov_size=cov_size, embedding_size=embedding_size,
region_size=region_size, sequence_length=sequence_length, multi_conv=muti_conv).to(device)
criterion = nn.CrossEntropyLoss().to(device)
optimizer = optim.Adam(textcnn.parameters(), lr=learning_rate)
if record:
writer = SummaryWriter()
all_step = 0
for i in tqdm(range(epoch)):
# 测试
right_num = 0
test_total_loss = 0
for j, (x_test, y_test) in enumerate(test_data_loader):
x_test = x_test.to(device)
y_test = y_test.to(device)
y_test_pre = textcnn(x_test)
test_total_loss += criterion(y_test_pre, y_test)
right_num += sum(torch.argmax(y_test_pre, dim=1) == y_test)
acc_test = float(right_num / test_len)
print()
print('第{}个epoch训练开始，测试集准确率{}，总损失{}'.format(i + 1, acc_test, test_total_loss))
if record:
writer.add_scalar('test_accuracy', acc_test, i+1)
writer.add_scalar('test_total_loss', test_total_loss, i + 1)
# 训练
for j, (x_train, y_train) in enumerate(train_data_loader):
x_train = x_train.to(device)
y_train = y_train.to(device)
y_train_pre = textcnn(x_train)
loss = criterion(y_train_pre, y_train)
if record:
if all_step % 1000 == 0:
writer.add_scalar('train_loss', loss, all_step)
all_step += 1
optimizer.zero_grad()
loss.backward()
optimizer.step()
if record:
torch.save(textcnn, 'model_1\\epoch_{}.pth'.format(i + 1))