{"id":29834,"date":"2024-10-28T03:00:27","date_gmt":"2024-10-28T03:00:27","guid":{"rendered":"http:\/\/atmokpo.com\/w\/?p=29834"},"modified":"2024-11-26T06:51:11","modified_gmt":"2024-11-26T06:51:11","slug":"%ed%8c%8c%ec%9d%b4%ed%86%a0%ec%b9%98%eb%a5%bc-%ed%99%9c%ec%9a%a9%ed%95%9c-gan-%eb%94%a5%eb%9f%ac%eb%8b%9d-%eb%af%b8%ec%88%a0-%ec%a0%84%ec%8b%9c%ed%9a%8c","status":"publish","type":"post","link":"https:\/\/atmokpo.com\/w\/29834\/","title":{"rendered":"\ud30c\uc774\ud1a0\uce58\ub97c \ud65c\uc6a9\ud55c GAN \ub525\ub7ec\ub2dd, \ubbf8\uc220 \uc804\uc2dc\ud68c"},"content":{"rendered":"

<\/p>\n

\uc11c\ub860<\/h2>\n

\n GAN(Generative Adversarial Network)\uc740 \uc0dd\uc131 \ubaa8\ub378\uc758 \uc77c\uc885\uc73c\ub85c, \ub450 \uac1c\uc758 \uc2e0\uacbd\ub9dd\uc774 \uc0c1\ud638\uc791\uc6a9\ud558\uc5ec \uc0c8\ub85c\uc6b4 \ub370\uc774\ud130\ub97c \uc0dd\uc131\ud558\ub294 \ubc29\ubc95\uc785\ub2c8\ub2e4. GAN\uc740 \uc8fc\ub85c \uc774\ubbf8\uc9c0 \uc0dd\uc131, \ud14d\uc2a4\ud2b8 \uc0dd\uc131, \uc74c\uc545 \uc0dd\uc131 \ub4f1 \ub2e4\uc591\ud55c \ubd84\uc57c\uc5d0\uc11c \ud65c\uc6a9\ub429\ub2c8\ub2e4. \ubcf8 \uae00\uc5d0\uc11c\ub294 \ud30c\uc774\ud1a0\uce58(PyTorch) \ub77c\uc774\ube0c\ub7ec\ub9ac\ub97c \ud65c\uc6a9\ud558\uc5ec GAN\uc744 \uad6c\ud604\ud558\uace0, \uc774\ub97c \ud1b5\ud574 \ubbf8\uc220 \uc804\uc2dc\ud68c\uc5d0\uc11c \uc0ac\uc6a9\ud560 \uc218 \uc788\ub294 \uc608\uc220 \uc791\ud488\uc744 \uc0dd\uc131\ud558\ub294 \uacfc\uc815\uc744 \uc0c1\uc138\ud788 \uc124\uba85\ud558\uaca0\uc2b5\ub2c8\ub2e4.\n <\/p>\n

1. GAN\uc758 \uae30\ubcf8 \uac1c\ub150<\/h2>\n

\n GAN\uc740 \ub450 \uac1c\uc758 \ubaa8\ub378, \uc989 \uc0dd\uc131\uc790(Generator)\uc640 \ud310\ubcc4\uc790(Discriminator)\ub85c \uad6c\uc131\ub429\ub2c8\ub2e4. \uc0dd\uc131\uc790\ub294 \ubb34\uc791\uc704 \uc785\ub825\uc744 \ubc14\ud0d5\uc73c\ub85c \ub370\uc774\ud130\ub97c \uc0dd\uc131\ud558\uba70, \ud310\ubcc4\uc790\ub294 \uc8fc\uc5b4\uc9c4 \ub370\uc774\ud130\uac00 \uc2e4\uc81c \ub370\uc774\ud130\uc778\uc9c0 \uc0dd\uc131\ub41c \ub370\uc774\ud130\uc778\uc9c0\ub97c \ud310\ubcc4\ud569\ub2c8\ub2e4. GAN\uc758 \ud559\uc2b5 \uacfc\uc815\uc740 \ub2e4\uc74c\uacfc \uac19\uc2b5\ub2c8\ub2e4.\n <\/p>\n

    \n
  1. \uc0dd\uc131\uc790\ub294 \ubb34\uc791\uc704 \ub178\uc774\uc988\ub97c \uc785\ub825\ubc1b\uc544 \uac00\uc9dc \uc774\ubbf8\uc9c0\ub97c \uc0dd\uc131\ud569\ub2c8\ub2e4.<\/li>\n
  2. \ud310\ubcc4\uc790\ub294 \uc0dd\uc131\ub41c \uc774\ubbf8\uc9c0\uc640 \uc2e4\uc81c \uc774\ubbf8\uc9c0\ub97c \ube44\uad50\ud569\ub2c8\ub2e4.<\/li>\n
  3. \ud310\ubcc4\uc790\uac00 \uac00\uc9dc \uc774\ubbf8\uc9c0\ub97c \uc2e4\uc81c \uc774\ubbf8\uc9c0\ub85c \uc798\ubabb \ud310\ub2e8\ud560\uc218\ub85d \uc0dd\uc131\uc790\ub294 \ub354 \ub098\uc740 \uc774\ubbf8\uc9c0\ub97c \uc0dd\uc131\ud558\ub3c4\ub85d \ud559\uc2b5\ud569\ub2c8\ub2e4.<\/li>\n<\/ol>\n

    \n \uc774 \uacfc\uc815\uc740 \uc0dd\uc131\uc790\uac00 \ub108\ubb34 \uac00\ubcbc\uc6b4 \uac00\uc9dc \uc774\ubbf8\uc9c0\ub97c \uc0dd\uc131\ud560 \uc218 \uc5c6\ub3c4\ub85d \ud310\ubcc4\uc790\uac00 \uc810\uc810 \ub354 \uc815\uad50\ud574\uc9c0\ub294 \ubaa8\uc2b5\uc744 \ub9cc\ub4e4\uc5b4\ub0c5\ub2c8\ub2e4.\n <\/p>\n

    2. \ud30c\uc774\ud1a0\uce58 \uc124\uce58\ud558\uae30<\/h2>\n

    \n GAN\uc744 \uad6c\ud604\ud558\uae30 \uc704\ud574 \uba3c\uc800 \ud30c\uc774\ud1a0\uce58\ub97c \uc124\uce58\ud574\uc57c \ud569\ub2c8\ub2e4. \ub2e4\uc74c\uc758 \uba85\ub839\uc5b4\ub97c \ud1b5\ud574 \uc124\uce58\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n

    \n    \n    pip install torch torchvision\n    <\/code>\n    <\/pre>\n
    3. \ub370\uc774\ud130 \uc900\ube44<\/h2>\n
    \n        \ubbf8\uc220 \uc791\ud488 \uc774\ubbf8\uc9c0\ub97c \uc0dd\uc131\ud558\uae30 \uc704\ud574\uc11c\ub294 \ub370\uc774\ud130\uc14b\uc774 \ud544\uc694\ud569\ub2c8\ub2e4. \uc774\ubc88 \uc2e4\uc2b5\uc5d0\uc11c\ub294 CIFAR-10 \ub370\uc774\ud130\uc14b\uc744 \uc0ac\uc6a9\ud558\uaca0\uc2b5\ub2c8\ub2e4. \uc774 \ub370\uc774\ud130\uc14b\uc740 10\uac00\uc9c0 \ud074\ub798\uc2a4\uc758 \uc774\ubbf8\uc9c0\ub85c \uad6c\uc131\ub418\uc5b4 \uc788\uc73c\uba70, \uadf8\ub9bc \ubbf8\uc220\uacfc \uad00\ub828\ub41c \ub370\uc774\ud130\uc14b\uc744 \uad6c\ucd95\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \ud30c\uc774\ud1a0\uce58 \ub0b4\uc7a5 \ud568\uc218\ub97c \uc0ac\uc6a9\ud574 \uc190\uc27d\uac8c \uc0ac\uc6a9\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n    <\/p>\n
    3.1 CIFAR-10 \ub370\uc774\ud130\uc14b<\/h3>\n
    \n    \n    import torchvision.transforms as transforms\n    import torchvision.datasets as datasets\n    from torch.utils.data import DataLoader\n\n    # \ub370\uc774\ud130 \uc804\ucc98\ub9ac\n    transform = transforms.Compose([\n        transforms.Resize(64),\n        transforms.CenterCrop(64),\n        transforms.ToTensor(),\n        transforms.Normalize([0.5, 0.5, 0.5], [0.5, 0.5, 0.5]),\n    ])\n\n    # CIFAR-10 \ub370\uc774\ud130\uc14b \ub85c\ub4dc\n    dataset = datasets.CIFAR10(root='.\/data', train=True, download=True, transform=transform)\n    dataloader = DataLoader(dataset, batch_size=128, shuffle=True, num_workers=2)\n    <\/code>\n    <\/pre>\n
    4. GAN \ubaa8\ub378 \uad6c\ud604\ud558\uae30<\/h2>\n
    \n        GAN \ubaa8\ub378\uc740 \ub450 \uac1c\uc758 \uc2e0\uacbd\ub9dd\uc744 \uc0ac\uc6a9\ud558\uc5ec \uc0dd\uc131\uc790\uc640 \ud310\ubcc4\uc790\ub97c \uc815\uc758\ud574\uc57c \ud569\ub2c8\ub2e4. \uc0dd\uc131\uc790\ub294 \uc784\uc758\uc758 \ub178\uc774\uc988\ub97c \uae30\ubc18\uc73c\ub85c \uc774\ubbf8\uc9c0\ub97c \uc0dd\uc131\ud558\uace0, \ud310\ubcc4\uc790\ub294 \uc774\ubbf8\uc9c0\uc758 \uc9c4\uc704\ub97c \ud310\ubcc4\ud569\ub2c8\ub2e4.\n    <\/p>\n
    4.1 \uc0dd\uc131\uc790 \ubaa8\ub378<\/h3>\n
    \n    \n    import torch\n    import torch.nn as nn\n\n    class Generator(nn.Module):\n        def __init__(self):\n            super(Generator, self).__init__()\n            self.model = nn.Sequential(\n                nn.Linear(100, 256),\n                nn.ReLU(),\n                nn.Linear(256, 512),\n                nn.ReLU(),\n                nn.Linear(512, 1024),\n                nn.ReLU(),\n                nn.Linear(1024, 3*64*64),\n                nn.Tanh(),\n            )\n\n        def forward(self, z):\n            z = self.model(z)\n            return z.view(-1, 3, 64, 64)\n    <\/code>\n    <\/pre>\n
    4.2 \ud310\ubcc4\uc790 \ubaa8\ub378<\/h3>\n
    \n    \n    class Discriminator(nn.Module):\n        def __init__(self):\n            super(Discriminator, self).__init__()\n            self.model = nn.Sequential(\n                nn.Linear(3*64*64, 512),\n                nn.LeakyReLU(0.2),\n                nn.Linear(512, 256),\n                nn.LeakyReLU(0.2),\n                nn.Linear(256, 1),\n                nn.Sigmoid(),\n            )\n\n        def forward(self, img):\n            img = img.view(-1, 3*64*64)\n            return self.model(img)\n    <\/code>\n    <\/pre>\n
    5. \uc190\uc2e4 \ud568\uc218\uc640 \uc635\ud2f0\ub9c8\uc774\uc800 \uc124\uc815\ud558\uae30<\/h2>\n
    \n        GAN\uc758 \ud559\uc2b5\uc744 \uc704\ud574 \uc190\uc2e4 \ud568\uc218\uc640 \uc635\ud2f0\ub9c8\uc774\uc800\ub97c \uc124\uc815\ud574\uc57c \ud569\ub2c8\ub2e4. \uc77c\ubc18\uc801\uc73c\ub85c \uc774\uc9c4 \uad50\ucc28 \uc5d4\ud2b8\ub85c\ud53c \uc190\uc2e4(Binary Cross Entropy Loss)\uc744 \uc0ac\uc6a9\ud558\uba70, Adam \ucd5c\uc801\ud654\ub97c \uc0ac\uc6a9\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n    <\/p>\n
    \n    \n    import torch.optim as optim\n\n    generator = Generator()\n    discriminator = Discriminator()\n\n    criterion = nn.BCELoss()\n    optimizer_G = optim.Adam(generator.parameters(), lr=0.0002, betas=(0.5, 0.999))\n    optimizer_D = optim.Adam(discriminator.parameters(), lr=0.0002, betas=(0.5, 0.999))\n    <\/code>\n    <\/pre>\n
    6. GAN \ud6c8\ub828\ud558\uae30<\/h2>\n
    \n        GAN\uc758 \ud6c8\ub828 \uacfc\uc815\uc740 \uc0dd\uc131\uc790\uc640 \ud310\ubcc4\uc790\uac00 \ubc88\uac08\uc544 \uac00\uba70 \ud559\uc2b5\ud558\ub294 \ubc29\uc2dd\uc785\ub2c8\ub2e4. \uc544\ub798\uc758 \ucf54\ub4dc\ub294 GAN\uc744 \ud6c8\ub828\ud558\uae30 \uc704\ud55c \uc8fc \ub8e8\ud504\ub97c \ub098\ud0c0\ub0c5\ub2c8\ub2e4.\n    <\/p>\n
    \n    \n    import numpy as np\n    import matplotlib.pyplot as plt\n\n    def train_gan(num_epochs):\n        for epoch in range(num_epochs):\n            for i, (imgs, _) in enumerate(dataloader):\n                # \uc9c4\uc9dc \uc774\ubbf8\uc9c0 \ub808\uc774\ube14: 1\n                real_labels = torch.ones(imgs.size(0), 1)\n                # \uac00\uc9dc \uc774\ubbf8\uc9c0 \ub808\uc774\ube14: 0\n                fake_labels = torch.zeros(imgs.size(0), 1)\n\n                # \ud310\ubcc4\uc790 \ud559\uc2b5\n                optimizer_D.zero_grad()\n                outputs = discriminator(imgs)\n                d_loss_real = criterion(outputs, real_labels)\n                d_loss_real.backward()\n\n                z = torch.randn(imgs.size(0), 100)\n                fake_images = generator(z)\n                outputs = discriminator(fake_images.detach())\n                d_loss_fake = criterion(outputs, fake_labels)\n                d_loss_fake.backward()\n\n                optimizer_D.step()\n                d_loss = d_loss_real + d_loss_fake\n\n                # \uc0dd\uc131\uc790 \ud559\uc2b5\n                optimizer_G.zero_grad()\n                outputs = discriminator(fake_images)\n                g_loss = criterion(outputs, real_labels)\n                g_loss.backward()\n                optimizer_G.step()\n\n                if (i % 100 == 0):\n                    print(f'Epoch [{epoch}\/{num_epochs}], Step [{i}\/{len(dataloader)}], d_loss: {d_loss.item()}, g_loss: {g_loss.item()}')\n\n    train_gan(num_epochs=20)\n    <\/code>\n    <\/pre>\n
    7. \uacb0\uacfc \uc2dc\uac01\ud654\ud558\uae30<\/h2>\n
    \n        \ubaa8\ub378 \ud6c8\ub828\uc774 \ub05d\ub098\uba74 \uc0dd\uc131\ub41c \uc774\ubbf8\uc9c0\ub97c \uc2dc\uac01\ud654\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uc544\ub798\uc758 \ucf54\ub4dc\ub294 \uc0dd\uc131\ub41c \uc774\ubbf8\uc9c0\ub97c \uc800\uc7a5\ud558\uace0 \uc2dc\uac01\ud654\ud558\ub294 \uacfc\uc815\uc785\ub2c8\ub2e4.\n    <\/p>\n
    \n    \n    def plot_generated_images(num_images):\n        z = torch.randn(num_images, 100)\n        generated_images = generator(z).detach().numpy()\n        generated_images = (generated_images + 1) \/ 2  # (-1, 1) -> (0, 1)\n        \n        fig, axes = plt.subplots(1, num_images, figsize=(15, 5))\n        for i in range(num_images):\n            axes[i].imshow(generated_images[i].transpose(1, 2, 0))\n            axes[i].axis('off')\n        plt.show()\n\n    plot_generated_images(10)\n    <\/code>\n    <\/pre>\n
    8. \uacb0\ub860<\/h2>\n
    \n        \ubcf8 \uae00\uc5d0\uc11c\ub294 \ud30c\uc774\ud1a0\uce58\ub97c \ud65c\uc6a9\ud55c GAN \ub525\ub7ec\ub2dd\uc758 \uae30\ubcf8 \uac1c\ub150 \ubc0f \uad6c\ud604 \uacfc\uc815\uc744 \uc0b4\ud3b4\ubcf4\uc558\uc2b5\ub2c8\ub2e4. GAN\uc744 \uc0ac\uc6a9\ud558\uc5ec \ubbf8\uc220 \uc791\ud488\uacfc \uac19\uc740 \uc0c8\ub85c\uc6b4 \uc774\ubbf8\uc9c0\ub97c \uc0dd\uc131\ud560 \uc218 \uc788\uc73c\uba70, \uc774\ub97c \ubbf8\uc220 \uc804\uc2dc\ud68c\uc640 \uac19\uc740 \ud589\uc0ac\uc5d0 \ud65c\uc6a9\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uc55e\uc73c\ub85c GAN \uae30\uc220\uc740 \ub354\uc6b1 \ubc1c\uc804\ud560 \uac83\uc73c\ub85c \uae30\ub300\ub418\uba70, \ub2e4\uc591\ud55c \uc870\ud615\uc801 \uac00\ub2a5\uc131\uc744 \uc81c\uacf5\ud560 \uac83\uc785\ub2c8\ub2e4.\n    <\/p>\n
    9. \ucd94\uac00 \uc790\ub8cc<\/h2>\n
    \n        \uc544\ub798\ub294 GAN \ubc0f \uad00\ub828 \uc8fc\uc81c\uc5d0 \ub300\ud55c \ucd94\uac00 \uc790\ub8cc\uc785\ub2c8\ub2e4.\n    <\/p>\n