{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.14","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"none","dataSources":[{"sourceId":21755,"databundleVersionId":1475600,"sourceType":"competition"}],"dockerImageVersionId":30804,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import torch\nimport torch.nn as nn\nimport torch.optim as optim\nimport torchvision.transforms as transforms\nfrom torchvision.utils import save_image\nfrom torch.utils.data import DataLoader, Dataset\nfrom PIL import Image\nimport os\nimport zipfile\n\n# Hyperparameters\nbatch_size = 64\nlr = 0.0002\nnz = 100  # Dimensi noise\nnum_epochs = 2  # Untuk pengujian cepat\nimage_size = 256  # Ukuran gambar 256x256 sesuai aturan kompetisi\nchannels_img = 3\nfeatures_g = 64\nfeatures_d = 64\nreal_label = 1.0\nfake_label = 0.0\n\n# Path dataset\nmonet_path = '/kaggle/input/gan-getting-started/monet_jpg'\nassert os.path.exists(monet_path), \"Path dataset Monet tidak ditemukan!\"\n","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"execution":{"iopub.status.busy":"2024-12-09T09:07:58.789224Z","iopub.execute_input":"2024-12-09T09:07:58.789631Z","iopub.status.idle":"2024-12-09T09:07:58.796776Z","shell.execute_reply.started":"2024-12-09T09:07:58.789595Z","shell.execute_reply":"2024-12-09T09:07:58.795789Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Transformasi dataset\ntransform = transforms.Compose([\n    transforms.Resize((image_size, image_size)),  # Ukuran gambar 256x256\n    transforms.ToTensor(),\n    transforms.Normalize([0.5] * channels_img, [0.5] * channels_img)\n])\n\n# Dataset custom\nclass MonetDataset(Dataset):\n    def __init__(self, root_dir, transform=None):\n        self.root_dir = root_dir\n        self.transform = transform\n        self.image_paths = [os.path.join(root_dir, f) for f in os.listdir(root_dir) if f.endswith(('.jpg', '.png'))]\n\n    def __len__(self):\n        return len(self.image_paths)\n\n    def __getitem__(self, idx):\n        img_path = self.image_paths[idx]\n        image = Image.open(img_path).convert('RGB')\n        if self.transform:\n            image = self.transform(image)\n        return image\n\n# Load dataset dan DataLoader\ndataset = MonetDataset(root_dir=monet_path, transform=transform)\ndataloader = DataLoader(dataset, batch_size=batch_size, shuffle=True)\n\nprint(f\"Jumlah batch: {len(dataloader)}\")\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-09T09:08:10.829525Z","iopub.execute_input":"2024-12-09T09:08:10.830463Z","iopub.status.idle":"2024-12-09T09:08:10.841999Z","shell.execute_reply.started":"2024-12-09T09:08:10.830408Z","shell.execute_reply":"2024-12-09T09:08:10.840740Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class Generator(nn.Module):\n    def __init__(self, nz, channels_img, features_g):\n        super(Generator, self).__init__()\n        self.gen = nn.Sequential(\n            self.block(nz, features_g * 8, 4, 1, 0),  # 4x4\n            self.block(features_g * 8, features_g * 4, 4, 2, 1),  # 8x8\n            self.block(features_g * 4, features_g * 2, 4, 2, 1),  # 16x16\n            self.block(features_g * 2, features_g, 4, 2, 1),  # 32x32\n            nn.ConvTranspose2d(features_g, channels_img, kernel_size=4, stride=2, padding=1),\n            nn.Tanh()  # Output image [-1, 1]\n        )\n\n    def block(self, in_channels, out_channels, kernel_size, stride, padding):\n        return nn.Sequential(\n            nn.ConvTranspose2d(in_channels, out_channels, kernel_size, stride, padding, bias=False),\n            nn.BatchNorm2d(out_channels),\n            nn.ReLU()\n        )\n\n    def forward(self, x):\n        return self.gen(x)\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-09T09:08:21.497298Z","iopub.execute_input":"2024-12-09T09:08:21.498297Z","iopub.status.idle":"2024-12-09T09:08:21.505573Z","shell.execute_reply.started":"2024-12-09T09:08:21.498253Z","shell.execute_reply":"2024-12-09T09:08:21.504373Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"class Discriminator(nn.Module):\n    def __init__(self, channels_img, features_d):\n        super(Discriminator, self).__init__()\n        self.disc = nn.Sequential(\n            self.block(channels_img, features_d, 4, 2, 1),  # 32x32\n            self.block(features_d, features_d * 2, 4, 2, 1),  # 16x16\n            self.block(features_d * 2, features_d * 4, 4, 2, 1),  # 8x8\n            self.block(features_d * 4, features_d * 8, 4, 2, 1),  # 4x4\n            nn.Conv2d(features_d * 8, 1, kernel_size=4, stride=1, padding=0),\n            nn.Sigmoid()  # Output single scalar [0, 1]\n        )\n\n    def block(self, in_channels, out_channels, kernel_size, stride, padding):\n        return nn.Sequential(\n            nn.Conv2d(in_channels, out_channels, kernel_size, stride, padding, bias=False),\n            nn.BatchNorm2d(out_channels),\n            nn.LeakyReLU(0.2)\n        )\n\n    def forward(self, x):\n        return self.disc(x)\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-09T09:08:29.154207Z","iopub.execute_input":"2024-12-09T09:08:29.154608Z","iopub.status.idle":"2024-12-09T09:08:29.162267Z","shell.execute_reply.started":"2024-12-09T09:08:29.154573Z","shell.execute_reply":"2024-12-09T09:08:29.161078Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"device = 'cuda' if torch.cuda.is_available() else 'cpu'\n\n# Inisialisasi model\ngen = Generator(nz, channels_img, features_g).to(device)\ndisc = Discriminator(channels_img, features_d).to(device)\n\n# Optimizer dan Loss\nopt_gen = optim.Adam(gen.parameters(), lr=lr, betas=(0.5, 0.999))\nopt_disc = optim.Adam(disc.parameters(), lr=lr, betas=(0.5, 0.999))\ncriterion = nn.BCELoss()\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-09T09:08:46.752793Z","iopub.execute_input":"2024-12-09T09:08:46.753487Z","iopub.status.idle":"2024-12-09T09:08:46.812298Z","shell.execute_reply.started":"2024-12-09T09:08:46.753443Z","shell.execute_reply":"2024-12-09T09:08:46.811247Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Pelatihan GAN\nfor epoch in range(num_epochs):\n    for batch_idx, real in enumerate(dataloader):\n        real = real.to(device)\n        batch_size = real.size(0)\n        \n        noise = torch.randn(batch_size, nz, 1, 1, device=device)\n        fake = gen(noise)\n\n        # Latih Discriminator\n        disc_real = disc(real).view(-1)\n        loss_disc_real = criterion(disc_real, torch.ones_like(disc_real) * real_label)\n        disc_fake = disc(fake.detach()).view(-1)\n        loss_disc_fake = criterion(disc_fake, torch.zeros_like(disc_fake))\n        loss_disc = (loss_disc_real + loss_disc_fake) / 2\n        opt_disc.zero_grad()\n        loss_disc.backward()\n        opt_disc.step()\n\n        # Latih Generator\n        output = disc(fake).view(-1)\n        loss_gen = criterion(output, torch.ones_like(output) * real_label)\n        opt_gen.zero_grad()\n        loss_gen.backward()\n        opt_gen.step()\n\n        if batch_idx % 100 == 0:\n            print(f\"Epoch [{epoch}/{num_epochs}] Batch {batch_idx}/{len(dataloader)} Loss D: {loss_disc:.4f}, Loss G: {loss_gen:.4f}\")\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-09T09:08:56.387692Z","iopub.execute_input":"2024-12-09T09:08:56.388198Z","iopub.status.idle":"2024-12-09T09:10:29.760637Z","shell.execute_reply.started":"2024-12-09T09:08:56.388151Z","shell.execute_reply":"2024-12-09T09:10:29.759450Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"zip_filename = \"/kaggle/working/images.zip\"\n\n# Membuka ZIP untuk menyimpan gambar\nwith zipfile.ZipFile(zip_filename, 'w') as zipf:\n    for i in range(7000):  # Pastikan jumlah gambar 7,000–10,000\n        noise = torch.randn(1, nz, 1, 1, device=device)\n        fake_img = gen(noise).detach().cpu()\n        img_path = f\"image_{i}.jpg\"\n        \n        # Simpan gambar langsung ke ZIP\n        save_image(fake_img, img_path, normalize=True)\n        zipf.write(img_path)\n        os.remove(img_path)  # Hapus file lokal setelah ditulis ke ZIP\n\nprint(f\"File ZIP '{zip_filename}' berhasil dibuat.\")\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-09T09:10:29.762399Z","iopub.execute_input":"2024-12-09T09:10:29.762774Z","iopub.status.idle":"2024-12-09T09:11:51.899743Z","shell.execute_reply.started":"2024-12-09T09:10:29.762740Z","shell.execute_reply":"2024-12-09T09:11:51.898388Z"}},"outputs":[],"execution_count":null}]}