{"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":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        print(os.path.join(dirname, filename))\n\n# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"execution":{"iopub.status.busy":"2024-12-17T14:39:27.737113Z","iopub.execute_input":"2024-12-17T14:39:27.737524Z","iopub.status.idle":"2024-12-17T14:39:34.199216Z","shell.execute_reply.started":"2024-12-17T14:39:27.737488Z","shell.execute_reply":"2024-12-17T14:39:34.198075Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import os\nimport numpy as np\nimport time\nimport gc\nimport matplotlib.pyplot as plt\nfrom tensorflow.keras.models import Sequential, Model\nfrom tensorflow.keras.layers import Dense, Reshape, Flatten, Conv2DTranspose, Conv2D, BatchNormalization, ReLU, LeakyReLU, Dropout, Input\nfrom tensorflow.keras.optimizers import Adam\nfrom tensorflow.keras.preprocessing.image import load_img, img_to_array\nfrom PIL import Image\nimport zipfile\nfrom glob import glob","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-17T14:39:48.533585Z","iopub.execute_input":"2024-12-17T14:39:48.534176Z","iopub.status.idle":"2024-12-17T14:40:02.411850Z","shell.execute_reply.started":"2024-12-17T14:39:48.534138Z","shell.execute_reply":"2024-12-17T14:40:02.410250Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Load and preprocess images\ndef load_images(image_paths, size=(128, 128)):\n    images = []\n    for path in image_paths:\n        img = load_img(path, target_size=size)\n        img = (img_to_array(img) - 127.5) / 127.5\n        images.append(img)\n    return np.array(images)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-17T14:40:08.151665Z","iopub.execute_input":"2024-12-17T14:40:08.152296Z","iopub.status.idle":"2024-12-17T14:40:08.158979Z","shell.execute_reply.started":"2024-12-17T14:40:08.152258Z","shell.execute_reply":"2024-12-17T14:40:08.157501Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Define the generator\ndef build_generator(latent_dim):\n    model = Sequential([\n        Dense(8 * 8 * 512, input_shape=(latent_dim,), use_bias=False),\n        BatchNormalization(), ReLU(), Reshape((8, 8, 512)),\n        Conv2DTranspose(256, 5, strides=2, padding='same', use_bias=False),\n        BatchNormalization(), ReLU(),\n        Conv2DTranspose(128, 5, strides=2, padding='same', use_bias=False),\n        BatchNormalization(), ReLU(),\n        Conv2DTranspose(64, 5, strides=2, padding='same', use_bias=False),\n        BatchNormalization(), ReLU(),\n        Conv2DTranspose(32, 5, strides=2, padding='same', use_bias=False),\n        BatchNormalization(), ReLU(),\n        Conv2DTranspose(3, 5, activation='tanh', padding='same')\n    ])\n    return model","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-17T14:40:18.337185Z","iopub.execute_input":"2024-12-17T14:40:18.337638Z","iopub.status.idle":"2024-12-17T14:40:18.345617Z","shell.execute_reply.started":"2024-12-17T14:40:18.337599Z","shell.execute_reply":"2024-12-17T14:40:18.344286Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Define the discriminator\ndef build_discriminator(input_shape):\n    model = Sequential([\n        Conv2D(64, 4, strides=2, padding='same', input_shape=input_shape),\n        LeakyReLU(0.2), Dropout(0.3),\n        Conv2D(128, 4, strides=2, padding='same'), BatchNormalization(), LeakyReLU(0.2), Dropout(0.3),\n        Conv2D(256, 4, strides=2, padding='same'), BatchNormalization(), LeakyReLU(0.2), Dropout(0.3),\n        Conv2D(512, 4, strides=2, padding='same'), BatchNormalization(), LeakyReLU(0.2), Dropout(0.3),\n        Flatten(), Dense(1, activation='sigmoid')\n    ])\n    return model","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-17T14:40:27.539382Z","iopub.execute_input":"2024-12-17T14:40:27.539962Z","iopub.status.idle":"2024-12-17T14:40:27.548476Z","shell.execute_reply.started":"2024-12-17T14:40:27.539916Z","shell.execute_reply":"2024-12-17T14:40:27.546869Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# GAN training function\ndef train_gan(generator, discriminator, gan, images, epochs, batch_size, latent_dim):\n    half_batch = batch_size // 2\n    d_losses, g_losses = [], []\n\n    for epoch in range(epochs):\n        d_loss_epoch, g_loss_epoch = [], []\n        for _ in range(images.shape[0] // batch_size):\n            real_imgs = images[np.random.randint(0, images.shape[0], half_batch)]\n            noise = np.random.normal(0, 1, (half_batch, latent_dim))\n            fake_imgs = generator.predict(noise)\n\n            d_loss_real = discriminator.train_on_batch(real_imgs, np.ones((half_batch, 1)) * 0.9)\n            d_loss_fake = discriminator.train_on_batch(fake_imgs, np.zeros((half_batch, 1)))\n            d_loss_epoch.append(0.5 * (d_loss_real[0] + d_loss_fake[0]))\n\n            noise = np.random.normal(0, 1, (batch_size, latent_dim))\n            g_loss_epoch.append(gan.train_on_batch(noise, np.ones((batch_size, 1))))\n\n        d_losses.append(np.mean(d_loss_epoch))\n        g_losses.append(np.mean(g_loss_epoch))\n        print(f\"Epoch {epoch+1}/{epochs} | D Loss: {d_losses[-1]:.4f} | G Loss: {g_losses[-1]:.4f}\")\n        gc.collect()\n\n    return d_losses, g_losses","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-17T14:40:38.049341Z","iopub.execute_input":"2024-12-17T14:40:38.049941Z","iopub.status.idle":"2024-12-17T14:40:38.060881Z","shell.execute_reply.started":"2024-12-17T14:40:38.049899Z","shell.execute_reply":"2024-12-17T14:40:38.059505Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Save generated images\ndef generate_and_save_images(generator, latent_dim, total_images, output_dir='images'):\n    os.makedirs(output_dir, exist_ok=True)\n    for i in range(total_images):\n        noise = np.random.normal(0, 1, (1, latent_dim))\n        img = ((generator.predict(noise)[0] * 127.5) + 127.5).astype(np.uint8)\n        Image.fromarray(img).save(f\"{output_dir}/image_{i+1}.png\")\n    print(\"Image generation completed.\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-17T14:40:50.350469Z","iopub.execute_input":"2024-12-17T14:40:50.350894Z","iopub.status.idle":"2024-12-17T14:40:50.358250Z","shell.execute_reply.started":"2024-12-17T14:40:50.350859Z","shell.execute_reply":"2024-12-17T14:40:50.357004Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Main process\nmonet_paths = glob('../input/gan-getting-started/monet_jpg/*.jpg')\nimages = load_images(monet_paths)\n\nlatent_dim = 100\ngenerator = build_generator(latent_dim)\ndiscriminator = build_discriminator((128, 128, 3))\n\ngenerator_optimizer = Adam(0.0002, 0.5)\ndiscriminator_optimizer = Adam(0.0002, 0.5)\ndiscriminator.compile(loss='binary_crossentropy', optimizer=discriminator_optimizer, metrics=['accuracy'])","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-17T14:41:06.018658Z","iopub.execute_input":"2024-12-17T14:41:06.019055Z","iopub.status.idle":"2024-12-17T14:41:07.794319Z","shell.execute_reply.started":"2024-12-17T14:41:06.019020Z","shell.execute_reply":"2024-12-17T14:41:07.793330Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# GAN model\ndiscriminator.trainable = False\ngan_input = Input(shape=(latent_dim,))\ngan_output = discriminator(generator(gan_input))\ngan = Model(gan_input, gan_output)\ngan.compile(loss='binary_crossentropy', optimizer=generator_optimizer)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-17T14:41:21.730944Z","iopub.execute_input":"2024-12-17T14:41:21.731337Z","iopub.status.idle":"2024-12-17T14:41:21.747578Z","shell.execute_reply.started":"2024-12-17T14:41:21.731304Z","shell.execute_reply":"2024-12-17T14:41:21.746052Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Train and save results\nepochs, batch_size = 10, 4\nd_losses, g_losses = train_gan(generator, discriminator, gan, images, epochs, batch_size, latent_dim)\n\nplt.plot(d_losses, label='Discriminator Loss')\nplt.plot(g_losses, label='Generator Loss')\nplt.legend(), plt.title(\"Loss Over Epochs\"), plt.show()\n\ngenerate_and_save_images(generator, latent_dim, total_images=7000)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-17T14:41:33.220949Z","iopub.execute_input":"2024-12-17T14:41:33.221328Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Zip generated images\nwith zipfile.ZipFile(\"images.zip\", 'w') as zf:\n    for f in os.listdir(\"images\"):\n        zf.write(f\"images/{f}\", f)\nprint(\"Images zipped successfully.\") ","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-17T15:44:47.087041Z","iopub.execute_input":"2024-12-17T15:44:47.087543Z","iopub.status.idle":"2024-12-17T15:44:48.454798Z","shell.execute_reply.started":"2024-12-17T15:44:47.087501Z","shell.execute_reply":"2024-12-17T15:44:48.453526Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}