/
test_misc.py
43 lines (34 loc) · 1.23 KB
/
test_misc.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
import torch
import torch.utils.data as udata
import torchvision
from torchvision import datasets as dsets, transforms
# x = torch.arange(0, 10).view(10, 1)
# xdata = udata.TensorDataset(x)
# xloader = udata.DataLoader(xdata, batch_size=3, drop_last=True, shuffle=True)
#
# xiter = iter(xloader)
# for j in range(10):
# try:
# print(next(xiter))
# except:
# xiter = iter(xloader)
# print(next(xiter))
size = 28
transformMnist = transforms.Compose([
transforms.Resize(size),
transforms.CenterCrop(size),
transforms.ToTensor(),
transforms.Normalize((0.5,), (0.5,)),
])
mnist = dsets.MNIST(root='~/github/data/mnist', train=True, transform=transformMnist, download=True)
mnist = udata.DataLoader(mnist, batch_size=64, drop_last=True, shuffle=True)
noise_weights = [0, 0.1, 0.5, 1., 2, 5]
for i, (x, y) in enumerate(mnist):
for nw in noise_weights:
noise = (torch.rand_like(x) * 2 - 1) * nw
xnoise = x + noise
torchvision.utils.save_image(noise, 'results/noise_%.1f.png' % nw, normalize=True, range=(-1, 1))
torchvision.utils.save_image(xnoise, 'results/noisy_%.1f.png' % nw, normalize=True) # , range=(-1.5, 1.5))
# end for
break
# end for