style_transfer.py

import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from PIL import Image
import torchvision.transforms as transforms
import torchvision.models as models
import torchvision
import argparse
import copy

def image_loader(image_name):
    image = Image.open(image_name)
    image = loader(image).unsqueeze(0)
    return image.to(device, torch.float)

class ContentLoss(nn.Module):

    def __init__(self, target,):
        super(ContentLoss, self).__init__()
        self.target = target.detach()

    def forward(self, input):
        self.loss = F.mse_loss(input, self.target)
        return input
def gram_matrix(input):
    batch, channel, height, width = input.size() 
    features = input.view(batch * channel, height * width)
    G = torch.mm(features, features.t()) 
    return G.div(batch *  channel * height *  width)

class StyleLoss(nn.Module):

    def __init__(self, target_feature):
        super(StyleLoss, self).__init__()
        self.target = gram_matrix(target_feature).detach()

    def forward(self, input):
        G = gram_matrix(input)
        self.loss = F.mse_loss(G, self.target)
        return input

class Normalization(nn.Module):
    def __init__(self):
        super(Normalization, self).__init__()
        mean = torch.tensor([0.485, 0.456, 0.406]).to(device)
        std = torch.tensor([0.229, 0.224, 0.225]).to(device)
        self.mean = torch.tensor(mean).view(-1, 1, 1)
        self.std = torch.tensor(std).view(-1, 1, 1)

    def forward(self, img):
        return (img - self.mean) / self.std


def get_style_model_and_losses(cnn,style_img, content_img):
    normalization = Normalization().to(device)
    # Desired layers to calculate content and style loss
    content_layers = ['conv_4']
    style_layers = ['conv_1', 'conv_2', 'conv_3', 'conv_4', 'conv_5']

    content_losses = []
    style_losses = []

    model = nn.Sequential(normalization)

    i = 0 
    for layer in cnn.children():
        if isinstance(layer, nn.Conv2d):
            i += 1
            name = 'conv_{}'.format(i)
        elif isinstance(layer, nn.ReLU):
            name = 'relu_{}'.format(i)
            layer = nn.ReLU(inplace=False)
        elif isinstance(layer, nn.MaxPool2d):
            name = 'pool_{}'.format(i)
        elif isinstance(layer, nn.BatchNorm2d):
            name = 'bn_{}'.format(i)
        else:
            raise RuntimeError('Unrecognized layer: {}'.format(layer.__class__.__name__))

        model.add_module(name, layer)

        # If the current layer is in desired content layers,
        # add it to the the model
        if name in content_layers:
            target = model(content_img).detach()
            content_loss = ContentLoss(target)
            model.add_module("content_loss_{}".format(i), content_loss)
            content_losses.append(content_loss)

        # If the current layer is in desired style layers,
        # add it to the the sequential model
        if name in style_layers:
            target_feature = model(style_img).detach()
            style_loss = StyleLoss(target_feature)
            model.add_module("style_loss_{}".format(i), style_loss)
            style_losses.append(style_loss)

    # Trim the layers after last content or style layer
    for i in range(len(model) - 1, -1, -1):
        if isinstance(model[i], ContentLoss) or isinstance(model[i], StyleLoss):
            break
    model = model[:(i + 1)]
    return model, style_losses, content_losses


def get_input_optimizer(input_img):
    optimizer = optim.LBFGS([input_img])
    return optimizer

def main(cnn,content_img, style_img, input_img, num_steps,style_weight, content_weight):

    model, style_losses, content_losses = get_style_model_and_losses(cnn,style_img, content_img)

    # Optimize the input image not the network
    input_img.requires_grad_(True)
    model.requires_grad_(False)
    optimizer = get_input_optimizer(input_img)

    print('#####Optimizing Image#####')
    run = [0]
    while run[0] <= num_steps:

        def closure():
            # correct the values of updated input image
            with torch.no_grad():
                input_img.clamp_(0, 1)

            optimizer.zero_grad()
            model(input_img)
            style_score = 0
            content_score = 0

            for sl in style_losses:
                style_score += sl.loss
            for cl in content_losses:
                content_score += cl.loss

            style_score *= style_weight
            content_score *= content_weight

            loss = style_score + content_score
            loss.backward()

            run[0] += 1
            if run[0] % 50 == 0:
                print("Epoch {}:".format(run))
                print('Style Loss : {:4f} Content Loss: {:4f}'.format(
                    style_score.item(), content_score.item()))
                print()

            return style_score + content_score

        optimizer.step(closure)

    with torch.no_grad():
        input_img.clamp_(0, 1)

    return input_img


if __name__ == "__main__":

    parser = argparse.ArgumentParser()
    parser.add_argument('-c', help = "content image", required = True)
    parser.add_argument('-s', help = "style image", required = True)
    parser.add_argument('-style_weight', help = "style weight", type = int, default = 1000000)
    parser.add_argument('-content_weight', help = "content weight",type = int,  default = 1)
    parser.add_argument('-steps', help = "number of steps",type = int,  default = 300)
    parser.add_argument('-save', help = "generated image name",type = str,required = True) 
    args = parser.parse_args()

    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    # Desired image size
    imsize = 512
    loader = transforms.Compose([
        transforms.Resize((imsize, imsize)),  
        transforms.ToTensor()])

    style_img = image_loader(args.s)
    content_img = image_loader(args.c)

    # Pretrained CNN model
    cnn = models.vgg19(pretrained=True).features.to(device).eval()

    input_img = content_img.clone()

    output = main(cnn, content_img, style_img, input_img,args.steps, args.style_weight, args.content_weight)

    torchvision.utils.save_image(output, args.save + ".png")
    print("Image Saved")