GAN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# GANs with MNIST"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import os,sys\n",
    "\n",
    "import tensorflow\n",
    "from tensorflow.keras.layers import Dense, Input, BatchNormalization,Dropout,LeakyReLU\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.optimizers import Adam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Data Loading\n",
    "mnist=tensorflow.keras.datasets.mnist\n",
    "(X_train,y_train),(X_test,y_test)=mnist.load_data()\n",
    "\n",
    "#scaling to between -1 and 1\n",
    "X_train=X_train/255.0\n",
    "X_test=X_test/255.0\n",
    "X_train=X_train*2-1\n",
    "X_test=X_test*2-1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Flatten the data\n",
    "X_train=X_train.reshape(-1,X_train.shape[1]*X_train.shape[2])\n",
    "X_test=X_test.reshape(-1,X_test.shape[1]*X_test.shape[2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(60000, 784)\n",
      "(10000, 784)\n",
      "784\n"
     ]
    }
   ],
   "source": [
    "D=X_train.shape[1]\n",
    "print(X_train.shape)\n",
    "print(X_test.shape)\n",
    "print(D)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Create Generator Architecture\n",
    "l_dim=100 #latemt dimensionality\n",
    "\n",
    "def build_gen(l_dim):\n",
    "    a=Input(shape=(l_dim,))\n",
    "    b=Dense(512,activation=LeakyReLU(alpha=0.2))(a)\n",
    "    b=BatchNormalization(momentum=0.7)(b)\n",
    "    b=Dense(1024,activation=LeakyReLU(alpha=0.2))(b)\n",
    "    b=BatchNormalization(momentum=0.7)(b)\n",
    "    b=Dense(1024,activation=LeakyReLU(alpha=0.2))(b)\n",
    "    b=BatchNormalization(momentum=0.7)(b)\n",
    "    b=Dense(D,activation='tanh')(b) #activation is tanh because images are scaled between -1 and 1\n",
    "    \n",
    "    model=Model(a,b)\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Create Discriminator Architecture\n",
    "\n",
    "def build_dis(image_size):\n",
    "    a=Input(shape=(image_size,))\n",
    "    b=Dense(512,activation=LeakyReLU(alpha=0.2))(a)\n",
    "    b=Dense(1024,activation=LeakyReLU(alpha=0.2))(b)\n",
    "    b=Dense(1,activation='sigmoid')(b)\n",
    "    \n",
    "    model=Model(a,b)\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Building and Compiling the ANNs\n",
    "\n",
    "dis=build_dis(D)\n",
    "dis.compile(optimizer=Adam(0.0001),loss='binary_crossentropy',metrics=['accuracy'])\n",
    "\n",
    "gen=build_gen(l_dim)\n",
    "\n",
    "z=Input(shape=(l_dim,))\n",
    "img=gen(z)\n",
    "dis.trainable=False\n",
    "pred=dis(img)\n",
    "comb_model=Model(z,pred)\n",
    "\n",
    "comb_model.compile(optimizer=Adam(0.0001),loss='binary_crossentropy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Training the GAN\n",
    "\n",
    "epochs=40000\n",
    "batch_size=16\n",
    "sample=1000\n",
    "\n",
    "ones=np.ones(batch_size)\n",
    "zeros=np.zeros(batch_size)\n",
    "\n",
    "dis_loss,gen_loss=[],[]\n",
    "\n",
    "if os.path.exists('gan_images')==False:\n",
    "    os.makedirs('gan_images')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Function to generate and store random samples \n",
    "\n",
    "def get_sample(epoch_no):\n",
    "    row,col=5,5 #25 images in a grid\n",
    "    noise= np.random.randn(row*col,l_dim)\n",
    "    imgs=gen.predict(noise)\n",
    "    \n",
    "    imgs= 0.5*imgs+0.5 #Scaling from the interval [-1,1] to [0,1]\n",
    "    \n",
    "    fig,axs=plt.subplots(row,col)\n",
    "    idx=0\n",
    "    for i in range(row):\n",
    "        for j in range(col):\n",
    "            axs[i,j].imshow(imgs[idx].reshape(28,28),cmap='gray') #As Height and Width of images are 28 pixels each\n",
    "            axs[i,j].axis('off')\n",
    "            idx+=1\n",
    "    fig.savefig('gan_images/%d.jpg' %epoch_no)\n",
    "    plt.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epoch: 1/40000,d_loss= 0.63, d_acc= 0.53, g_loss=  1.13\n"
     ]
    }
   ],
   "source": [
    "#Training loop\n",
    "\n",
    "\n",
    "for epoch in range(1):\n",
    "    \n",
    "    #Training the Discriminator\n",
    "    \n",
    "    #real images\n",
    "    k=np.random.randint(0,X_train.shape[0],batch_size)\n",
    "    real_img=X_train[k]\n",
    "    \n",
    "    #fake images\n",
    "    l=np.random.rand(batch_size,l_dim)\n",
    "    fake_img=gen.predict(l)\n",
    "    \n",
    "    #Train Discriminator\n",
    "    dis_loss_real,dis_acc_real=dis.train_on_batch(real_img,ones)\n",
    "    dis_loss_fake,dis_acc_fake=dis.train_on_batch(fake_img,zeros)\n",
    "    \n",
    "    d_loss=(dis_loss_real+dis_loss_fake)/2\n",
    "    dis_acc=(dis_acc_real+dis_acc_fake)/2\n",
    "    \n",
    "    #Training the Generator\n",
    "    \n",
    "    noise=np.random.rand(batch_size,l_dim)\n",
    "    g_loss=comb_model.train_on_batch(noise,ones)\n",
    "    \n",
    "    gen_loss.append(g_loss)\n",
    "    dis_loss.append(d_loss)\n",
    "    \n",
    "    if epoch%200==0:\n",
    "        print(f'epoch: {epoch+1}/{epochs},d_loss= {d_loss:.2f}, d_acc={dis_acc: 0.2f}, g_loss= {g_loss: 0.2f}')\n",
    "        \n",
    "    if epoch%sample==0:\n",
    "        get_sample(epoch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(dis_loss,label='discriminator')\n",
    "plt.plot(gen_loss,label='generator')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x2524d93a988>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "img=plt.imread('gan_images/39000.jpg')\n",
    "plt.imshow(img)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}