File size: 1,871 Bytes
476e0f0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import torch
import torch.nn as nn
import torch.nn.functional as F

class UpsampleBlock(nn.Module):
    def __init__(self, num_input_channels, num_output_channels):
        super(UpsampleBlock, self).__init__()
        self.pixel_shuffle = nn.PixelShuffle(2)
        self.conv = nn.Conv2d(num_input_channels // (2 * 2), num_output_channels, 3, stride=1, padding=1)
        self.relu = nn.ReLU()
        
    def forward(self, x):
        x = self.pixel_shuffle(x)
        x = self.conv(x)
        x = self.relu(x)
        return x
    
class AppearanceNetwork(nn.Module):
    def __init__(self, num_input_channels, num_output_channels):
        super(AppearanceNetwork, self).__init__()
        
        self.conv1 = nn.Conv2d(num_input_channels, 256, 3, stride=1, padding=1)
        self.up1 = UpsampleBlock(256, 128)
        self.up2 = UpsampleBlock(128, 64)
        self.up3 = UpsampleBlock(64, 32)
        self.up4 = UpsampleBlock(32, 16)
        
        self.conv2 = nn.Conv2d(16, 16, 3, stride=1, padding=1)
        self.conv3 = nn.Conv2d(16, num_output_channels, 3, stride=1, padding=1)
        self.relu = nn.ReLU()
        self.sigmoid = nn.Sigmoid()
        
    def forward(self, x):
        x = self.conv1(x)
        x = self.relu(x)
        x = self.up1(x)
        x = self.up2(x)
        x = self.up3(x)
        x = self.up4(x)
        # bilinear interpolation
        x = F.interpolate(x, scale_factor=2, mode='bilinear', align_corners=True)
        x = self.conv2(x)
        x = self.relu(x)
        x = self.conv3(x)
        x = self.sigmoid(x)
        return x


if __name__ == "__main__":
    H, W = 1200//32, 1600//32
    input_channels = 3 + 64
    output_channels = 3
    input = torch.randn(1, input_channels, H, W).cuda()
    model = AppearanceNetwork(input_channels, output_channels).cuda()
    
    output = model(input)
    print(output.shape)