Generative Adversarial Networks (GANs)\uc740 \ub450 \uac1c\uc758 \uc2e0\uacbd\ub9dd\uc774 \uacbd\uc7c1\ud558\uc5ec \uc0dd\uc131\ub41c \ub370\uc774\ud130\uc758 \ud488\uc9c8\uc744 \ud5a5\uc0c1\uc2dc\ud0a4\ub294 \ub525\ub7ec\ub2dd \ud504\ub808\uc784\uc6cc\ud06c\uc785\ub2c8\ub2e4. GAN\uc758 \uae30\ubcf8 \uad6c\uc870\ub294 \uc0dd\uc131\uae30(generator)\uc640 \ud310\ubcc4\uae30(discriminator)\ub85c \uad6c\uc131\ub418\uc5b4 \uc788\uc2b5\ub2c8\ub2e4. \uc0dd\uc131\uae30\ub294 \uc2e4\uc81c \ub370\uc774\ud130\uc640 \uc720\uc0ac\ud55c \ub370\uc774\ud130\ub97c \uc0dd\uc131\ud558\ub824\uace0 \ud558\uba70, \ud310\ubcc4\uae30\ub294 \uc0dd\uc131\ub41c \ub370\uc774\ud130\uac00 \uc2e4\uc81c \ub370\uc774\ud130\uc778\uc9c0 \uc0dd\uc131\ub41c \ub370\uc774\ud130\uc778\uc9c0\ub97c \uad6c\ubcc4\ud569\ub2c8\ub2e4. \uc774 \ub450 \ub124\ud2b8\uc6cc\ud06c\ub294 \uc11c\ub85c\uc758 \uc131\ub2a5\uc744 \ud5a5\uc0c1\uc2dc\ud0a4\uae30 \uc704\ud574 \uacbd\uc7c1\ud558\uba70, \uc774\ub97c \ud1b5\ud574 \uc810\uc810 \ub354 \uc0ac\uc2e4\uc801\uc778 \ub370\uc774\ud130\ub97c \uc0dd\uc131\ud558\uac8c \ub429\ub2c8\ub2e4.<\/p>\n
GAN\uc758 \uad6c\uc870\ub294 \ub2e4\uc74c\uacfc \uac19\uc774 \uad6c\uc131\ub429\ub2c8\ub2e4:<\/p>\n
GAN\uc740 \uc544\ub798\uc640 \uac19\uc740 \ub450 \ub2e8\uacc4\uc758 \ud559\uc2b5 \uacfc\uc815\uc744 \uac70\uce69\ub2c8\ub2e4:<\/p>\n
\uc774\ubc88 \uc139\uc158\uc5d0\uc11c\ub294 PyTorch\ub97c \uc0ac\uc6a9\ud558\uc5ec \uac04\ub2e8\ud55c GAN\uc744 \uad6c\ud604\ud574\ubcf4\uaca0\uc2b5\ub2c8\ub2e4.<\/p>\n
python\nimport torch\nimport torch.nn as nn\nimport torch.optim as optim\nimport torchvision\nimport torchvision.transforms as transforms\nimport matplotlib.pyplot as plt\n<\/code><\/pre>\n2.2 \uc0dd\uc131\uae30\uc640 \ud310\ubcc4\uae30 \uc815\uc758<\/h3>\n
GAN\uc5d0\uc11c \uc0dd\uc131\uae30\uc640 \ud310\ubcc4\uae30\uc758 \uad6c\uc870\ub97c \uc815\uc758\ud569\ub2c8\ub2e4.<\/p>\n
python\nclass Generator(nn.Module):\n def __init__(self):\n super(Generator, self).__init__()\n self.model = nn.Sequential(\n nn.Dense(128, input_size=100),\n nn.ReLU(),\n nn.Dense(256),\n nn.ReLU(),\n nn.Dense(512),\n nn.ReLU(),\n nn.Dense(1, activation='tanh') # Assume output is 1D data\n )\n \n def forward(self, z):\n return self.model(z)\n\nclass Discriminator(nn.Module):\n def __init__(self):\n super(Discriminator, self).__init__()\n self.model = nn.Sequential(\n nn.Dense(512, input_size=1), # 1D data input\n nn.LeakyReLU(0.2),\n nn.Dense(256),\n nn.LeakyReLU(0.2),\n nn.Dense(1, activation='sigmoid') # Binary output\n )\n \n def forward(self, x):\n return self.model(x)\n<\/code><\/pre>\n2.3 GAN\uc758 \ud6c8\ub828 \uacfc\uc815<\/h3>\n
\uc774\uc81c GAN\uc744 \ud6c8\ub828\uc2dc\ud0a4\ub294 \uacfc\uc815\uc744 \uac19\uc2b5\ub2c8\ub2e4.<\/p>\n
python\ndef train_gan(num_epochs=10000, batch_size=64, learning_rate=0.0002):\n transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.5,), (0.5,))])\n dataset = torchvision.datasets.MNIST(root='.\/data', train=True, download=True, transform=transform)\n dataloader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, shuffle=True)\n\n generator = Generator()\n discriminator = Discriminator()\n criterion = nn.BCELoss()\n optimizer_g = optim.Adam(generator.parameters(), lr=learning_rate)\n optimizer_d = optim.Adam(discriminator.parameters(), lr=learning_rate)\n\n for epoch in range(num_epochs):\n for real_data, _ in dataloader:\n real_data = real_data.view(-1, 1).to(torch.float32)\n batch_size = real_data.size(0)\n\n # Train Discriminator\n optimizer_d.zero_grad()\n z = torch.randn(batch_size, 100)\n fake_data = generator(z).detach()\n real_label = torch.ones(batch_size, 1)\n fake_label = torch.zeros(batch_size, 1)\n output_real = discriminator(real_data)\n output_fake = discriminator(fake_data)\n loss_d = criterion(output_real, real_label) + criterion(output_fake, fake_label)\n loss_d.backward()\n optimizer_d.step()\n\n # Train Generator\n optimizer_g.zero_grad()\n z = torch.randn(batch_size, 100)\n fake_data = generator(z)\n output = discriminator(fake_data)\n loss_g = criterion(output, real_label)\n loss_g.backward()\n optimizer_g.step()\n\n if epoch % 1000 == 0:\n print(f'Epoch [{epoch}\/{num_epochs}], Loss D: {loss_d.item()}, Loss G: {loss_g.item()}')\n<\/code><\/pre>\n3. \uc6d4\ub4dc \ubaa8\ub378 \uad6c\uc870<\/h2>\n
\uc6d4\ub4dc \ubaa8\ub378\uc740 \ud658\uacbd\uc758 \ubaa8\ub378\uc744 \ud559\uc2b5\ud558\uace0, \uadf8 \ubaa8\ub378\uc744 \uc0ac\uc6a9\ud558\uc5ec \ub2e4\uc591\ud55c \uc2dc\ub098\ub9ac\uc624\ub97c \uc2dc\ubbac\ub808\uc774\uc158\ud558\uc5ec \ucd5c\uc801\uc758 \ud589\ub3d9\uc744 \ud559\uc2b5\ud558\ub294\ub370 \uc0ac\uc6a9\ub418\ub294 \uad6c\uc870\uc785\ub2c8\ub2e4. \uc774\ub294 \uac15\ud654 \ud559\uc2b5\uacfc \uc0dd\uc131 \ubaa8\ub378\uc758 \uc870\ud569\uc73c\ub85c \ubcfc \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
3.1 \uc6d4\ub4dc \ubaa8\ub378\uc758 \uad6c\uc131 \uc694\uc18c<\/h3>\n
\uc6d4\ub4dc \ubaa8\ub378\uc740 \ub2e4\uc74c \uc138 \uac00\uc9c0 \uae30\ubcf8 \uad6c\uc131 \uc694\uc18c\ub85c \uc774\ub8e8\uc5b4\uc838 \uc788\uc2b5\ub2c8\ub2e4:<\/p>\n
\n- \ube44\uc8fc\uc5bc \ubaa8\ub378(Visual Model)<\/strong>: \ud658\uacbd\uc758 \uc2dc\uac01\uc801 \uc0c1\ud0dc\ub97c \ubaa8\ub378\ub9c1\ud569\ub2c8\ub2e4.<\/li>\n
- \ub2e4\uc774\ub098\ubbf9 \ubaa8\ub378(Dynamic Model)<\/strong>: \uc0c1\ud0dc\uc5d0\uc11c \uc0c1\ud0dc\ub85c\uc758 \uc804\uc774\ub97c \ubaa8\ub378\ub9c1\ud569\ub2c8\ub2e4.<\/li>\n
- \ud589\ub3d9 \uc815\ucc45(Policy)<\/strong>: \uc2dc\ubbac\ub808\uc774\uc158 \uacb0\uacfc\ub97c \uae30\ubc18\uc73c\ub85c \ucd5c\uc801\uc758 \ud589\ub3d9\uc744 \uacb0\uc815\ud569\ub2c8\ub2e4.<\/li>\n<\/ul>\n
3.2 \uc6d4\ub4dc \ubaa8\ub378\uc758 PyTorch \uad6c\ud604<\/h3>\n
\ub2e4\uc74c\uc73c\ub85c, \uc6d4\ub4dc \ubaa8\ub378\uc758 \uac04\ub2e8\ud55c \uc608\uc81c\ub97c \uad6c\ud604\ud574\ubcf4\uaca0\uc2b5\ub2c8\ub2e4.<\/p>\n
python\nclass VisualModel(nn.Module):\n def __init__(self):\n super(VisualModel, self).__init__()\n self.model = nn.Sequential(\n nn.Linear(784, 128),\n nn.ReLU(),\n nn.Linear(128, 64),\n nn.ReLU(),\n nn.Linear(64, 32)\n )\n\n def forward(self, x):\n return self.model(x)\n\nclass DynamicModel(nn.Module):\n def __init__(self):\n super(DynamicModel, self).__init__()\n self.model = nn.Sequential(\n nn.Linear(32 + 10, 64), # \uc0c1\ud0dc + \ud589\ub3d9\n nn.ReLU(),\n nn.Linear(64, 32)\n )\n\n def forward(self, state, action):\n return self.model(torch.cat([state, action], dim=1))\n\nclass Policy(nn.Module):\n def __init__(self):\n super(Policy, self).__init__()\n self.model = nn.Sequential(\n nn.Linear(32, 64),\n nn.ReLU(),\n nn.Linear(64, 10) # 10\uac1c\uc758 \ud589\ub3d9\n )\n\n def forward(self, state):\n return self.model(state)\n<\/code><\/pre>\n3.3 \uc6d4\ub4dc \ubaa8\ub378 \ud6c8\ub828<\/h3>\n
\uac01 \ubaa8\ub378\uc744 \ud6c8\ub828\ud558\uc5ec \uc0c1\ud0dc\uc640 \ud589\ub3d9\uc758 \uad00\uacc4\ub97c \ud559\uc2b5\ud569\ub2c8\ub2e4. \uc774\ub97c \ud1b5\ud574 \ub2e4\uc591\ud55c \uc2dc\ubbac\ub808\uc774\uc158\uc744 \ud1b5\ud574 \ud589\ub3d9 \uc815\ucc45\uc744 \ud559\uc2b5\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
4. \uacb0\ub860<\/h2>\n
\uc5ec\uae30\uc11c \uc6b0\ub9ac\ub294 GAN\uacfc \uc6d4\ub4dc \ubaa8\ub378\uc758 \uae30\ubcf8 \uc6d0\ub9ac\ub97c \uc124\uba85\ud558\uace0, \uc774\ub97c PyTorch\ub85c \uad6c\ud604\ud558\ub294 \ubc29\ubc95\uc5d0 \ub300\ud574 \uc0b4\ud3b4\ubcf4\uc558\uc2b5\ub2c8\ub2e4. \uc774\ub7ec\ud55c \uad6c\uc131 \uc694\uc18c\ub4e4\uc740 \ub2e4\uc591\ud55c \uba38\uc2e0\ub7ec\ub2dd \ubc0f \ub525\ub7ec\ub2dd \uc751\uc6a9 \ud504\ub85c\uadf8\ub7a8\uc5d0\uc11c \uc911\uc694\ud55c \uc5ed\ud560\uc744 \ud558\uace0 \uc788\uc2b5\ub2c8\ub2e4. GAN\uc740 \uc774\ubbf8\uc9c0 \uc0dd\uc131, \uc6d4\ub4dc \ubaa8\ub378\uc740 \uc2dc\ubbac\ub808\uc774\uc158\uacfc \uc815\ucc45 \ud559\uc2b5\uc5d0 \uc801\ud569\ud569\ub2c8\ub2e4. \uc774\ub7ec\ud55c \uae30\ubc95\ub4e4\uc744 \ud1b5\ud574 \ub354\uc6b1 \uc815\uad50\ud55c \ubaa8\ub378\ub9c1\uacfc \ub370\uc774\ud130 \uc0dd\uc131\uc774 \uac00\ub2a5\ud574\uc9d1\ub2c8\ub2e4.<\/p>\n
5. \ucc38\uace0 \uc790\ub8cc<\/h2>\n