File size: 4,264 Bytes
3369069
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
import torch
import torch.nn as nn
import torch.optim as optim
import torchvision.transforms as transforms
import torchvision.models as models
from PIL import Image
import torchvision.transforms.functional as TF

# 🚀 Device configuration
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

# 🔧 Preprocessing
transform = transforms.Compose([
    transforms.Resize((512, 512)),
    transforms.ToTensor()
])

def load_image(img):
    image = img.convert("RGB")
    return transform(image).unsqueeze(0).to(device)

# 🎯 Loss modules
class Normalization(nn.Module):
    def __init__(self, mean, std):
        super().__init__()
        self.mean = mean.view(-1, 1, 1)
        self.std = std.view(-1, 1, 1)
    def forward(self, img):
        return (img - self.mean) / self.std

class ContentLoss(nn.Module):
    def __init__(self, target):
        super().__init__()
        self.target = target.detach()
        self.loss = 0
    def forward(self, input):
        self.loss = nn.functional.mse_loss(input, self.target)
        return input

def gram_matrix(input):
    b, c, h, w = input.size()
    features = input.view(c, h * w)
    G = torch.mm(features, features.t())
    return G.div(c * h * w)

class StyleLoss(nn.Module):
    def __init__(self, target_feature):
        super().__init__()
        self.target = gram_matrix(target_feature).detach()
        self.loss = 0
    def forward(self, input):
        G = gram_matrix(input)
        self.loss = nn.functional.mse_loss(G, self.target)
        return input

# 🧬 Model builder
def get_model_losses(cnn, norm_mean, norm_std, style_img, content_img):
    normalization = Normalization(norm_mean, norm_std).to(device)
    model = nn.Sequential(normalization)

    content_losses = []
    style_losses = []

    i = 0
    for layer in cnn.children():
        name = None
        if isinstance(layer, nn.Conv2d):
            i += 1
            name = f"conv_{i}"
        elif isinstance(layer, nn.ReLU):
            name = f"relu_{i}"
            layer = nn.ReLU(inplace=False)
        elif isinstance(layer, nn.MaxPool2d):
            name = f"pool_{i}"
        elif isinstance(layer, nn.BatchNorm2d):
            name = f"bn_{i}"
        if name:
            model.add_module(name, layer)
            if name == "conv_4":
                target = model(content_img).detach()
                content_loss = ContentLoss(target)
                model.add_module(f"content_loss_{i}", content_loss)
                content_losses.append(content_loss)
            if name in ["conv_1", "conv_2", "conv_3", "conv_4", "conv_5"]:
                target_feature = model(style_img).detach()
                style_loss = StyleLoss(target_feature)
                model.add_module(f"style_loss_{i}", style_loss)
                style_losses.append(style_loss)

    for j in range(len(model) - 1, -1, -1):
        if isinstance(model[j], ContentLoss) or isinstance(model[j], StyleLoss):
            break

    return model[:j + 1], style_losses, content_losses

# ✨ Stylization pipeline
def run_nst(content_pil, style_pil, steps=300):
    content = load_image(content_pil)
    style = load_image(style_pil)
    input_img = content.clone().requires_grad_(True)

    cnn = models.vgg19(pretrained=True).features.to(device).eval()
    norm_mean = torch.tensor([0.485, 0.456, 0.406]).to(device)
    norm_std = torch.tensor([0.229, 0.224, 0.225]).to(device)

    model, style_losses, content_losses = get_model_losses(
        cnn, norm_mean, norm_std, style, content
    )

    optimizer = optim.LBFGS([input_img])
    run = [0]

    while run[0] <= steps:
        def closure():
            input_img.data.clamp_(0, 1)
            optimizer.zero_grad()
            model(input_img)

            style_score = sum(sl.loss for sl in style_losses)
            content_score = sum(cl.loss for cl in content_losses)

            loss = content_score + 1e6 * style_score
            loss.backward()

            run[0] += 1
            return loss
        optimizer.step(closure)

    output = input_img.clone().detach().cpu().squeeze(0)
    return TF.to_pil_image(output)