-
Notifications
You must be signed in to change notification settings - Fork 0
/
trainae.py
96 lines (58 loc) · 2.2 KB
/
trainae.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
import torch
import numpy as np
from ae import aeNet, aeDataset
from torch.utils.data import DataLoader
import param
import time
import os
from torch.utils.tensorboard import SummaryWriter
files = os.listdir("airplaneData")
allfile = []
for id, file in enumerate(files):
datapath = os.path.join("airplaneData", file)
allfile.append(datapath)
x = np.array(allfile)
train_set, test_set = np.split(x, [int(len(x)*0.8)]) #60%训练集、30%测试集、10%验证集
trainset = aeDataset(train_set)
trainloader = DataLoader(trainset, batch_size=param.batch_size, shuffle=True)
testset = aeDataset(test_set)
testloader = DataLoader(testset, batch_size=param.batch_size, shuffle=True)
net = aeNet().to(param.device)
optimizer = torch.optim.Adam(net.parameters(), lr=1e-6, betas=param.beta)
def loss(gt, fake):
return torch.mean((gt - fake) ** 2)
def saveM(net, opt, PATH):
torch.save({
'G_state_dict': net.en.state_dict(),
'D_state_dict': net.de.state_dict(),
'D_opt_state_dict': opt.state_dict(),
}, PATH)
writer = SummaryWriter('tensorboard_aesave')
print("start training!")
for epoch in range(param.epochs):
epoch_start_time = time.time()
train_loss = 0.0
val_loss = 0.0
net.train()
for i, data in enumerate(trainloader):
data = data.to(param.device)
optimizer.zero_grad()
out = net(data)
batch_loss = loss(out, data)
batch_loss.backward()
optimizer.step()
train_loss += batch_loss.item()
net.eval()
with torch.no_grad():
for i, data in enumerate(trainloader):
data = data.to(param.device)
optimizer.zero_grad()
out = net(data)
batch_loss = loss(out, data)
val_loss += batch_loss.item()
writer.add_scalars('Loss', {'Train': train_loss / len(trainset)}, epoch)
writer.add_scalars('Loss', {'Test': val_loss / len(testset)}, epoch)
if epoch % 10 == 0:
saveM(net, optimizer, "AE_M/"+str(epoch).zfill(3)+".pth")
print('[%03d/%03d] %2.2f sec(s) Train Loss: %3.6f | Val loss: %3.6f' % \
(epoch + 1, param.epochs, time.time() - epoch_start_time, train_loss / len(trainset), val_loss / len(testset)))