docs: add conv example for gan

Author: marcpinet

examples/models-usages/generation/gan-image-generation/gan-mnist-convolutional.ipynb

@@ -0,0 +1,171 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import random\n",
+    "import numpy as np\n",
+    "import matplotlib.pyplot as plt\n",
+    "import os\n",
+    "\n",
+    "from PIL import Image\n",
+    "from keras.datasets import mnist\n",
+    "from IPython.display import Image as IPImage\n",
+    "\n",
+    "from neuralnetlib.preprocessing import one_hot_encode\n",
+    "from neuralnetlib.models import Sequential, GAN\n",
+    "from neuralnetlib.layers import Input, Dense, Conv2D, Reshape, Flatten, UpSampling2D"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Load the MNIST dataset\n",
+    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
+    "n_classes = np.unique(y_train).shape[0]\n",
+    "\n",
+    "# Reshape images to include channel dimension\n",
+    "x_train = x_train.reshape(x_train.shape[0], 28, 28, 1)\n",
+    "x_test = x_test.reshape(x_test.shape[0], 28, 28, 1)\n",
+    "\n",
+    "# Normalize pixel values\n",
+    "x_train = x_train.astype('float32') / 255\n",
+    "x_test = x_test.astype('float32') / 255\n",
+    "\n",
+    "# Labels to categorical\n",
+    "y_train = one_hot_encode(y_train, n_classes)\n",
+    "y_test = one_hot_encode(y_test, n_classes)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "i = random.randint(0, len(x_train) - 1)\n",
+    "plt.imshow(x_train[i].reshape(28,28), cmap='gray')\n",
+    "plt.title('Class: ' + str(np.argmax(y_train[i])))\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "noise_dim = 32\n",
+    "\n",
+    "generator = Sequential()\n",
+    "generator.add(Input(noise_dim))\n",
+    "generator.add(Dense(7 * 7 * 128))\n",
+    "generator.add(Reshape((7, 7, 128)))\n",
+    "generator.add(UpSampling2D(size=(2, 2)))  # 14x14\n",
+    "generator.add(Conv2D(64, kernel_size=3, padding='same', activation='relu'))\n",
+    "generator.add(UpSampling2D(size=(2, 2)))  # 28x28\n",
+    "generator.add(Conv2D(32, kernel_size=3, padding='same', activation='relu'))\n",
+    "generator.add(Conv2D(1, kernel_size=3, padding='same', activation='sigmoid'))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "discriminator = Sequential()\n",
+    "discriminator.add(Input((28, 28, 1)))\n",
+    "discriminator.add(Conv2D(32, kernel_size=3, strides=2, padding='same', activation='relu'))  # 14x14\n",
+    "discriminator.add(Conv2D(64, kernel_size=3, strides=2, padding='same', activation='relu'))  # 7x7\n",
+    "discriminator.add(Flatten())\n",
+    "discriminator.add(Dense(128, activation='relu'))\n",
+    "discriminator.add(Dense(1, activation='sigmoid'))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "gan = GAN(latent_dim=noise_dim)\n",
+    "\n",
+    "gan.compile(\n",
+    "    generator,\n",
+    "    discriminator,\n",
+    "    generator_optimizer='adam',\n",
+    "    discriminator_optimizer='adam',\n",
+    "    loss_function='bce',\n",
+    "    verbose=True\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "history = gan.fit(x_train,\n",
+    "                 epochs=40,\n",
+    "                 batch_size=128,\n",
+    "                 plot_generated=True,\n",
+    "                 )   "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "image_files = [f for f in os.listdir() if f.endswith('.png') and f.startswith('video')]\n",
+    "image_files.sort(key=lambda x: int(x.replace('video', '').replace('.png', '')))\n",
+    "\n",
+    "images = [Image.open(img) for img in image_files]\n",
+    "\n",
+    "if images:\n",
+    "    images[0].save('output.gif', save_all=True, append_images=images[1:], duration=100, loop=0)\n",
+    "\n",
+    "print(\"GIF 'output.gif' succesffuly created!\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "IPImage(filename=\"output.gif\")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.8"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}