-
Notifications
You must be signed in to change notification settings - Fork 0
/
train.py
59 lines (43 loc) · 1.97 KB
/
train.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
import torch
import torch.nn as nn
import torch.optim as optim
from utils import weights_init, label_smoothing
from models import Generator, Discriminator
from config import *
import numpy as np
import matplotlib.pyplot as plt
import torchvision.utils as vutils
def train(device, dataloader, generator, discriminator, optimizer_G, optimizer_D, adversarial_loss, n_epochs, latent_dim, n_classes):
losses = {"G": [], "D": []}
for epoch in range(n_epochs):
for i, (imgs, labels) in enumerate(dataloader):
smoothed_label_value = 0.75
fake_label_value = 0.25
valid = torch.full((imgs.size(0), 1), smoothed_label_value, device=device)
fake = torch.full((imgs.size(0), 1), fake_label_value, device=device)
real_imgs = imgs.to(device)
labels = labels.to(device)
z = torch.randn(imgs.size(0), latent_dim, device=device)
gen_labels = torch.randint(0, n_classes, (imgs.size(0),), device=device)
gen_imgs = generator(z, gen_labels)
optimizer_D.zero_grad()
real_pred = discriminator(real_imgs, labels)
d_real_loss = adversarial_loss(real_pred, valid)
fake_pred = discriminator(gen_imgs.detach(), gen_labels)
d_fake_loss = adversarial_loss(fake_pred, fake)
d_loss = 0.5 * (d_real_loss + d_fake_loss)
d_loss.backward()
optimizer_D.step()
optimizer_G.zero_grad()
validity = discriminator(gen_imgs, gen_labels)
g_loss = adversarial_loss(validity, valid)
g_loss.backward()
optimizer_G.step()
losses["G"].append(g_loss.item())
losses["D"].append(d_loss.item())
if i % 50 == 0:
print(
"[Epoch %d/%d] [Batch %d/%d] [D loss: %.3f] [G loss: %.3f]"
% (epoch, n_epochs, i, len(dataloader), d_loss.item(), g_loss.item())
)
return losses