DCGAN_on_digits.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DCGANs on MNIST data\n",
    "You can run this notebook on google colab.\n",
    "\n",
    "<div align=\"center\"><a href=\"https://colab.research.google.com/github/vstark21/DCGANs/blob/master/DCGAN_on_digits.ipynb\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\"></a></div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "NZBGwEZVD9dD"
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "import numpy as np\n",
    "from IPython import display\n",
    "import matplotlib.pyplot as plt\n",
    "import time"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Creating functions - \n",
    "* `generator_model` - returns a model to generate images.\n",
    "* `discriminator_model` - returns a model to discriminate original images and generated images.\n",
    "* `generator_loss` - loss to optimize generator model so that it produces images like original images.\n",
    "* `discriminator_loss` - loss to optimize discriminator model so that it discriminates original images and generated images correctly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "2OR0tNKADzem"
   },
   "outputs": [],
   "source": [
    "def generator_model():\n",
    "\n",
    "    model = keras.Sequential()\n",
    "\n",
    "    model.add(keras.layers.Dense(7*7*256, input_shape=[100, ]))\n",
    "    model.add(keras.layers.BatchNormalization())\n",
    "    model.add(keras.layers.ReLU())\n",
    "\n",
    "    model.add(keras.layers.Reshape((7, 7, 256)))\n",
    "\n",
    "    model.add(keras.layers.Conv2DTranspose(128, (5, 5), strides=(1, 1), padding=\"same\"))\n",
    "    model.add(keras.layers.BatchNormalization())\n",
    "    model.add(keras.layers.ReLU())\n",
    "\n",
    "    model.add(keras.layers.Conv2DTranspose(64, (5, 5), strides=(2, 2), padding=\"same\"))\n",
    "    model.add(keras.layers.BatchNormalization())\n",
    "    model.add(keras.layers.ReLU()),\n",
    "\n",
    "    model.add(keras.layers.Conv2DTranspose(1, (5, 5), strides=(2, 2), padding=\"same\", activation=\"tanh\"))\n",
    "\n",
    "    return model\n",
    "\n",
    "def discriminator_model():\n",
    "\n",
    "    model = keras.Sequential([\n",
    "                             keras.layers.Conv2D(64, (5, 5), strides=(2, 2), input_shape=(28, 28, 1), padding=\"same\"),\n",
    "                               keras.layers.Dropout(0.3),\n",
    "                             keras.layers.LeakyReLU(),\n",
    "\n",
    "                             keras.layers.Conv2D(128, (5, 5), strides=(2, 2), padding=\"same\"),\n",
    "                             keras.layers.Dropout(0.3),\n",
    "                             keras.layers.LeakyReLU(),\n",
    "                              \n",
    "                             keras.layers.Conv2D(256, (5, 5), strides=(2, 2), padding=\"same\"),\n",
    "                             keras.layers.Dropout(0.3),\n",
    "                             keras.layers.LeakyReLU(),\n",
    "\n",
    "                             keras.layers.Flatten(),\n",
    "                             keras.layers.Dense(1, activation=\"sigmoid\")\n",
    "    ])\n",
    "\n",
    "    return model\n",
    "\n",
    "def generator_loss(generated_output):\n",
    "\n",
    "    cross_entropy = keras.losses.BinaryCrossentropy()\n",
    "    loss = cross_entropy(tf.ones_like(generated_output), generated_output)\n",
    "    \n",
    "    return loss\n",
    "\n",
    "\n",
    "def discriminator_loss(original_output, generated_output):\n",
    "\n",
    "    cross_entropy = keras.losses.BinaryCrossentropy()\n",
    "    \n",
    "    real_loss =  cross_entropy(tf.ones_like(original_output), original_output)\n",
    "    gen_loss = cross_entropy(tf.zeros_like(generated_output), generated_output)\n",
    "\n",
    "    total_loss = real_loss + gen_loss\n",
    "\n",
    "    return total_loss\n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Loading data from `tf.keras.datasets.mnist` then reshaping, normalizing them and slicing them into minibatches.\n",
    "\n",
    "**NOTE :** \n",
    "> All the training images should be normalized to [-1, 1] beacuse all the discminator's inputs must have values which are in same range as generator model's output is computed using *tanh* so it returns values in range [-1, 1]."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Ck1cUG_rEDhu"
   },
   "outputs": [],
   "source": [
    "(train_images, train_labels), (_, _) = tf.keras.datasets.mnist.load_data()\n",
    "\n",
    "train_images = train_images.reshape(train_images.shape[0], 28, 28, 1).astype('float32')\n",
    "train_images = (train_images - 127.5) / 127.5 # Normalize the images to [-1, 1]\n",
    "\n",
    "BUFFER_SIZE = 60000\n",
    "BATCH_SIZE = 256\n",
    "\n",
    "# Batch and shuffle the data\n",
    "train_dataset = tf.data.Dataset.from_tensor_slices(train_images).shuffle(BUFFER_SIZE).batch(BATCH_SIZE)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Creating a function `generate_and_save_images` so that we can save images after each epoch and creating a constant `TEST_INPUT` so that we can always use it to generate images to save."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "oqhUi_CIEE3p"
   },
   "outputs": [],
   "source": [
    "TEST_INPUT = tf.random.normal((16, 100))\n",
    "def generate_and_save_images(model, epoch, test_input):\n",
    "\n",
    "    predictions = model(test_input, training=False)\n",
    "\n",
    "    fig = plt.figure(figsize=(4,4))\n",
    "    for i in range(predictions.shape[0]):\n",
    "        plt.subplot(4, 4, i+1)\n",
    "        plt.imshow(predictions[i, :, :, 0] * 127.5 + 127.5, cmap='gray')\n",
    "        plt.axis('off')\n",
    "\n",
    "    plt.savefig('image_at_epoch_{:04d}.png'.format(epoch))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Creating a function `train_step` to train our models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "HEi-DrsZEFkb"
   },
   "outputs": [],
   "source": [
    "@tf.function\n",
    "def train_step(batch):\n",
    "    \n",
    "    noise = tf.random.normal([BATCH_SIZE, GEN_INPUT_DIMS])\n",
    "\n",
    "    with tf.GradientTape() as gen_tape, tf.GradientTape() as disc_tape:\n",
    "\n",
    "        generated_images = generator(noise, training=True)\n",
    "            \n",
    "        original_output = discriminator(batch, training=True)\n",
    "        generated_output = discriminator(generated_images, training=True)\n",
    "\n",
    "        gen_loss = generator_loss(generated_output)\n",
    "        disc_loss = discriminator_loss(original_output, generated_output)\n",
    "\n",
    "    gen_gradients = gen_tape.gradient(gen_loss, generator.trainable_variables)\n",
    "    disc_gradients = disc_tape.gradient(disc_loss, discriminator.trainable_variables)\n",
    "\n",
    "    generator_optimizer.apply_gradients(zip(gen_gradients, generator.trainable_variables))\n",
    "    discriminator_optimizer.apply_gradients(zip(disc_gradients, discriminator.trainable_variables))\n",
    "\n",
    "    return gen_loss, disc_loss\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Creating a generator model to test whether its output matches our required output."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 265
    },
    "colab_type": "code",
    "id": "3WwWmBqQEBxZ",
    "outputId": "544d2baf-440b-42f6-e924-50402e9fabc4"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "gen_test = generator_model()\n",
    "image = gen_test(tf.random.normal([3, 100]))\n",
    "plt.imshow(image[1, :, :, 0] * 127.5 + 127.5, cmap=\"gray\")\n",
    "\n",
    "EPOCHS = 100\n",
    "GEN_INPUT_DIMS = 100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Creating models and optimizers and training them using **mini-batch Gradient Descent**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 265
    },
    "colab_type": "code",
    "id": "ypb9b8qaEHBi",
    "outputId": "f276ebe6-fcd7-40a8-d05a-145cf609e308"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 16 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 99 took 45.63103413581848 with gen_loss of 0.9433663487434387 and disc_loss of 1.323568344116211\n"
     ]
    }
   ],
   "source": [
    "generator = generator_model()\n",
    "discriminator = discriminator_model()\n",
    "generator_optimizer = keras.optimizers.Adam(0.0001)\n",
    "discriminator_optimizer = keras.optimizers.Adam(0.0001)\n",
    "\n",
    "for epoch in range(EPOCHS):\n",
    "\n",
    "    print(f\"In epoch : {epoch}\")\n",
    "    epoch_start = time.time()\n",
    "\n",
    "    for batch in train_dataset:\n",
    "\n",
    "        gen_loss, disc_loss = train_step(batch)\n",
    "        \n",
    "\n",
    "    display.clear_output(wait=True)\n",
    "    generate_and_save_images(generator, epoch, TEST_INPUT)\n",
    "    print(f\"Epoch {epoch} took {time.time() - epoch_start} with gen_loss of {gen_loss} and disc_loss of {disc_loss}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Creating a *gif* using all the images generated."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "qXwtR4ZfEJdH"
   },
   "outputs": [],
   "source": [
    "import os, imageio\n",
    "\n",
    "files = os.listdir()\n",
    "\n",
    "images = []\n",
    "\n",
    "for filename in files:\n",
    "\n",
    "    if filename[-3:] == \"png\":\n",
    "        images.append(imageio.imread(filename))\n",
    "\n",
    "imageio.mimsave(\"Generating_Digits.gif\", images)"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "Pro0.ipynb",
   "provenance": []
  },
  "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": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}