GAN-keras-mnist-MLP.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.datasets import mnist\n",
    "from keras.utils import np_utils\n",
    "from keras.models import Sequential, Model\n",
    "from keras.layers import Input, Dense, Dropout, Activation, Flatten\n",
    "from keras.layers.advanced_activations import LeakyReLU\n",
    "from keras.optimizers import Adam, RMSprop\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import random\n",
    "from tqdm import tqdm_notebook\n",
    "\n",
    "# Dataset of 60,000 28x28 grayscale images of the 10 digits, along with a test set of 10,000 images.\n",
    "(X_train, Y_train), (X_test, Y_test) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Preprocessing\n",
    "\n",
    "X_train = X_train.reshape(60000, 784)\n",
    "X_test = X_test.reshape(10000, 784)\n",
    "X_train = X_train.astype('float32')/255\n",
    "X_test = X_test.astype('float32')/255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Set the dimensions of the noise\n",
    "z_dim = 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Optimizer\n",
    "adam = Adam(lr=0.0002, beta_1=0.5)\n",
    "\n",
    "g = Sequential()\n",
    "g.add(Dense(256, input_dim=z_dim, activation=LeakyReLU(alpha=0.2)))\n",
    "g.add(Dense(512, activation=LeakyReLU(alpha=0.2)))\n",
    "g.add(Dense(1024, activation=LeakyReLU(alpha=0.2)))\n",
    "g.add(Dense(784, activation='sigmoid'))  # Values between 0 and 1\n",
    "g.compile(loss='binary_crossentropy', optimizer=adam, metrics=['accuracy'])\n",
    "\n",
    "d = Sequential()\n",
    "d.add(Dense(1024, input_dim=784, activation=LeakyReLU(alpha=0.2)))\n",
    "d.add(Dropout(0.3))\n",
    "d.add(Dense(512, activation=LeakyReLU(alpha=0.2)))\n",
    "d.add(Dropout(0.3))\n",
    "d.add(Dense(256, activation=LeakyReLU(alpha=0.2)))\n",
    "d.add(Dropout(0.3))\n",
    "d.add(Dense(1, activation='sigmoid'))  # Values between 0 and 1\n",
    "d.compile(loss='binary_crossentropy', optimizer=adam, metrics=['accuracy'])\n",
    "\n",
    "d.trainable = False\n",
    "inputs = Input(shape=(z_dim, ))\n",
    "hidden = g(inputs)\n",
    "output = d(hidden)\n",
    "gan = Model(inputs, output)\n",
    "gan.compile(loss='binary_crossentropy', optimizer=adam, metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "def plot_loss(losses):\n",
    "    \"\"\"\n",
    "    @losses.keys():\n",
    "        0: loss\n",
    "        1: accuracy\n",
    "    \"\"\"\n",
    "    d_loss = [v[0] for v in losses[\"D\"]]\n",
    "    g_loss = [v[0] for v in losses[\"G\"]]\n",
    "    #d_acc = [v[1] for v in losses[\"D\"]]\n",
    "    #g_acc = [v[1] for v in losses[\"G\"]]\n",
    "    \n",
    "    plt.figure(figsize=(10,8))\n",
    "    plt.plot(d_loss, label=\"Discriminator loss\")\n",
    "    plt.plot(g_loss, label=\"Generator loss\")\n",
    "    #plt.plot(d_acc, label=\"Discriminator accuracy\")\n",
    "    #plt.plot(g_acc, label=\"Generator accuracy\")\n",
    "    plt.xlabel('Epochs')\n",
    "    plt.ylabel('Loss')\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "    \n",
    "def plot_generated(n_ex=10, dim=(1, 10), figsize=(12, 2)):\n",
    "    noise = np.random.normal(0, 1, size=(n_ex, z_dim))\n",
    "    generated_images = g.predict(noise)\n",
    "    generated_images = generated_images.reshape(n_ex, 28, 28)\n",
    "\n",
    "    plt.figure(figsize=figsize)\n",
    "    for i in range(generated_images.shape[0]):\n",
    "        plt.subplot(dim[0], dim[1], i+1)\n",
    "        plt.imshow(generated_images[i], interpolation='nearest', cmap='gray_r')\n",
    "        plt.axis('off')\n",
    "    plt.tight_layout()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Set up a vector (dict) to store the losses\n",
    "losses = {\"D\":[], \"G\":[]}\n",
    "\n",
    "def train(epochs=1, plt_frq=1, BATCH_SIZE=128):\n",
    "    batchCount = int(X_train.shape[0] / BATCH_SIZE)\n",
    "    print('Epochs:', epochs)\n",
    "    print('Batch size:', BATCH_SIZE)\n",
    "    print('Batches per epoch:', batchCount)\n",
    "    \n",
    "    for e in tqdm_notebook(range(1, epochs+1)):\n",
    "        if e == 1 or e%plt_frq == 0:\n",
    "            print('-'*15, 'Epoch %d' % e, '-'*15)\n",
    "        for _ in range(batchCount):  # tqdm_notebook(range(batchCount), leave=False):\n",
    "            # Create a batch by drawing random index numbers from the training set\n",
    "            image_batch = X_train[np.random.randint(0, X_train.shape[0], size=BATCH_SIZE)]\n",
    "            # Create noise vectors for the generator\n",
    "            noise = np.random.normal(0, 1, size=(BATCH_SIZE, z_dim))\n",
    "            \n",
    "            # Generate the images from the noise\n",
    "            generated_images = g.predict(noise)\n",
    "            X = np.concatenate((image_batch, generated_images))\n",
    "            # Create labels\n",
    "            y = np.zeros(2*BATCH_SIZE)\n",
    "            y[:BATCH_SIZE] = 0.9  # One-sided label smoothing\n",
    "\n",
    "            # Train discriminator on generated images\n",
    "            d.trainable = True\n",
    "            d_loss = d.train_on_batch(X, y)\n",
    "\n",
    "            # Train generator\n",
    "            noise = np.random.normal(0, 1, size=(BATCH_SIZE, z_dim))\n",
    "            y2 = np.ones(BATCH_SIZE)\n",
    "            d.trainable = False\n",
    "            g_loss = gan.train_on_batch(noise, y2)\n",
    "\n",
    "        # Only store losses from final batch of epoch\n",
    "        losses[\"D\"].append(d_loss)\n",
    "        losses[\"G\"].append(g_loss)\n",
    "\n",
    "        # Update the plots\n",
    "        if e == 1 or e%plt_frq == 0:\n",
    "            plot_generated()\n",
    "    plot_loss(losses)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epochs: 200\n",
      "Batch size: 128\n",
      "Batches per epoch: 468\n"
     ]
    },
    {
     "data": {},
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Epoch 1 ---------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAABWCAYAAADSfTdFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20594737a58>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Epoch 20 ---------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAABWCAYAAADSfTdFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSZIU5H+V+ycQs0zHWwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2059486a5f8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Epoch 40 ---------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAABWCAYAAADSfTdFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20595afcc50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Epoch 60 ---------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAABWCAYAAADSfTdFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20595e41d30>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Epoch 80 ---------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAABWCAYAAADSfTdFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20595f0b6d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Epoch 100 ---------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAABWCAYAAADSfTdFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20595c94080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Epoch 120 ---------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAABWCAYAAADSfTdFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJv8DlXJxsHkrgaAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20595a45390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Epoch 140 ---------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAABWCAYAAADSfTdFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCfECy8zMzMzMLCFeYJmZmZmZmSXkf3Nfv3kGI1enAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2059496af98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Epoch 160 ---------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAABWCAYAAADSfTdFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmZmZmZlZQvyAZWZmZmZmlpD/A4hLliYjF9/HAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20595de06a0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Epoch 180 ---------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAABWCAYAAADSfTdFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x205947f7b70>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------- Epoch 200 ---------------\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1gAAABWCAYAAADSfTdFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkqSKOMGSJEmSpIo4wZIkSZKkivw/MAx3mBzO1V4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20595995da0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmEAAAHjCAYAAABrZcgFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nwoQQQgghbPD/AW84n19XLgC5AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x20596011978>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train(epochs=200, plt_frq=20, BATCH_SIZE=128)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    ""
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3.0
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}