Test_Keras_WGAN_8.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Some stuff from\n",
    "https://github.com/tdeboissiere/DeepLearningImplementations/blob/master/WassersteinGAN/src/model/train_WGAN.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%run 'Lib.ipynb'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "OUT_DIR = 'out_8e'\n",
    "if not os.path.isdir(OUT_DIR): os.makedirs(OUT_DIR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# use GPU 1\n",
    "os.environ[\"CUDA_DEVICE_ORDER\"] = \"PCI_BUS_ID\"\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"1\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from collections import defaultdict\n",
    "\n",
    "import keras.backend as K\n",
    "from keras.datasets import mnist\n",
    "from keras.layers import *\n",
    "from keras.models import *\n",
    "from keras.optimizers import *\n",
    "from keras.utils.generic_utils import Progbar"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "Z_SIZE = 200\n",
    "\n",
    "# batch and latent size taken from the paper\n",
    "EPOCHS = 50\n",
    "BATCH_SIZE = 64\n",
    "\n",
    "# how many times to train D per each G(D) update\n",
    "# 5 is suggested in https://arxiv.org/pdf/1701.07875.pdf\n",
    "D_TRAIN_ITERATIONS = 5\n",
    "\n",
    "# D weight clipping per https://arxiv.org/pdf/1701.07875.pdf\n",
    "D_WEIGHTS_LOWER_BOUND = -0.01\n",
    "D_WEIGHTS_UPPER_BOUND = 0.01\n",
    "\n",
    "# learning rate, per https://arxiv.org/pdf/1701.07875.pdf \n",
    "RMSPROP_LR = 0.00005"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "np.random.seed(777)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "K.set_image_dim_ordering('tf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def create_D():\n",
    "    input_image = Input(shape=(28, 28, 1), name='input_image')\n",
    "\n",
    "    x = Conv2D(32, (3, 3), padding='same', name='conv_1')(input_image)\n",
    "    x = LeakyReLU()(x)\n",
    "    x = MaxPool2D(pool_size=2)(x)\n",
    "    x = Dropout(0.3)(x)\n",
    "\n",
    "    x = Conv2D(64, (3, 3), padding='same', name='conv_2')(x)\n",
    "    x = MaxPool2D(pool_size=1)(x)\n",
    "    x = LeakyReLU()(x)\n",
    "    x = Dropout(0.3)(x)\n",
    "\n",
    "    x = Conv2D(128, (3, 3), padding='same', name='conv_3')(x)\n",
    "    x = MaxPool2D(pool_size=2)(x)\n",
    "    x = LeakyReLU()(x)\n",
    "    x = Dropout(0.3)(x)\n",
    "\n",
    "    x = Conv2D(256, (3, 3), padding='same', name='coonv_4')(x)\n",
    "    x = MaxPool2D(pool_size=1)(x)\n",
    "    x = LeakyReLU()(x)\n",
    "    x = Dropout(0.3)(x)\n",
    "\n",
    "    features = Flatten()(x)\n",
    "\n",
    "    output_is_fake = Dense(\n",
    "        1, activation='linear', name='output_is_fake')(features)\n",
    "\n",
    "    return Model(inputs=[input_image], outputs=[output_is_fake], name='D')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def create_G(Z_SIZE=Z_SIZE):\n",
    "    # latent var\n",
    "    input_z = Input(shape=(Z_SIZE, ), name='input_z')\n",
    "\n",
    "    # cnn part\n",
    "\n",
    "    x = Dense(512 * 7 * 7)(input_z)\n",
    "    x = LeakyReLU()(x)\n",
    "    x = Reshape((7, 7, 512))(x)\n",
    "\n",
    "    x = UpSampling2D(size=(2, 2))(x)\n",
    "    x = Conv2D(256, (5, 5), padding='same')(x)\n",
    "    x = LeakyReLU()(x)\n",
    "\n",
    "    x = UpSampling2D(size=(2, 2))(x)\n",
    "    x = Conv2D(128, (5, 5), padding='same')(x)\n",
    "    x = LeakyReLU()(x)\n",
    "\n",
    "    x = Conv2D(1, (2, 2), padding='same', activation='tanh', name='output_generated_image')(x)\n",
    "\n",
    "    return Model(inputs=[input_z], outputs=[x], name='G')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def wasserstein(y_true, y_pred):\n",
    "    return K.mean(y_true * y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "D = create_D()\n",
    "D.compile(optimizer=RMSprop(lr=RMSPROP_LR), loss=[wasserstein])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "input_z = Input(shape=(Z_SIZE, ), name='input_z_')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "G = create_G()\n",
    "G.compile(loss='mse', optimizer=RMSprop(lr=RMSPROP_LR))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# create combined D(G) model\n",
    "output_is_fake = D(G(inputs=[input_z]))\n",
    "combined = Model(inputs=[input_z], outputs=[output_is_fake])\n",
    "combined.compile(optimizer=RMSprop(lr=RMSPROP_LR), loss=[wasserstein])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f8c54667160>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfoAAAH0CAYAAADVH+85AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8c58ef5048>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 250,
       "width": 253
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# try generating image\n",
    "z = np.random.randn(Z_SIZE).reshape(1, Z_SIZE)\n",
    "img = G.predict([z])\n",
    "plt.imshow(img.reshape((28, 28)), cmap=plt.cm.gray)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# get our mnist data, and force it to be of shape (..., 28, 28, 1) with\n",
    "# range [-1., 1.]\n",
    "(X_train, _), (X_test, _) = mnist.load_data()\n",
    "X_train = (X_train.astype(np.float32) - 127.5) / 127.5\n",
    "X_train = np.expand_dims(X_train, axis=3)\n",
    "X_test = (X_test.astype(np.float32) - 127.5) / 127.5\n",
    "X_test = np.expand_dims(X_test, axis=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f8c541fa7b8>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfoAAAH0CAYAAADVH+85AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRjEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8c54637b38>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 250,
       "width": 253
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X_train[777].reshape((28, 28)), cmap=plt.cm.gray)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "train_num_samples, test_num_samples = X_train.shape[0], X_test.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def save_generated_samples(n):\n",
    "\n",
    "    # new batch of z's\n",
    "    zz = np.random.normal(0., 1., (100, Z_SIZE))\n",
    "\n",
    "    generated_images = G.predict([zz])\n",
    "\n",
    "    rr = []\n",
    "    for c in range(10):\n",
    "        rr.append(np.concatenate(generated_images[c * 10:(1 + c) * 10]).reshape(280, 28))\n",
    "    img = np.hstack(rr)\n",
    "\n",
    "    plt.imsave(OUT_DIR + '/generated@%d.png'%n, img, cmap=plt.cm.gray)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 1 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 1.1060 - D_loss_real: 382.3305 - D_loss_gen: 381.2245 - DG_loss: 365.9005\n",
      "Epoch 23 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 1.4853 - D_loss_real: 381.6798 - D_loss_gen: 380.1945 - DG_loss: 365.1466\n",
      "Epoch 24 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 1.9678 - D_loss_real: 378.5526 - D_loss_gen: 376.5848 - DG_loss: 363.0572\n",
      "Epoch 25 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 2.5914 - D_loss_real: 374.8673 - D_loss_gen: 372.2759 - DG_loss: 360.6996\n",
      "Epoch 26 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 2.5295 - D_loss_real: 374.2844 - D_loss_gen: 371.7549 - DG_loss: 358.2811\n",
      "Epoch 27 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 2.7184 - D_loss_real: 375.1380 - D_loss_gen: 372.4196 - DG_loss: 357.9604\n",
      "Epoch 28 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 2.7711 - D_loss_real: 374.2345 - D_loss_gen: 371.4634 - DG_loss: 356.5976\n",
      "Epoch 29 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 3.0085 - D_loss_real: 373.7412 - D_loss_gen: 370.7328 - DG_loss: 355.9372\n",
      "Epoch 30 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 3.1226 - D_loss_real: 372.7574 - D_loss_gen: 369.6348 - DG_loss: 354.2997\n",
      "Epoch 31 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 3.5433 - D_loss_real: 371.2190 - D_loss_gen: 367.6757 - DG_loss: 353.2438\n",
      "Epoch 32 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 3.9067 - D_loss_real: 369.7458 - D_loss_gen: 365.8391 - DG_loss: 352.3686\n",
      "Epoch 33 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 4.2464 - D_loss_real: 366.9052 - D_loss_gen: 362.6588 - DG_loss: 350.7409\n",
      "Epoch 34 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 4.4816 - D_loss_real: 366.5296 - D_loss_gen: 362.0480 - DG_loss: 348.4942\n",
      "Epoch 35 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 4.5656 - D_loss_real: 367.2499 - D_loss_gen: 362.6844 - DG_loss: 348.3450\n",
      "Epoch 36 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 4.6401 - D_loss_real: 366.5283 - D_loss_gen: 361.8882 - DG_loss: 347.3225\n",
      "Epoch 37 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 4.8487 - D_loss_real: 365.7809 - D_loss_gen: 360.9322 - DG_loss: 345.8857\n",
      "Epoch 38 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 4.9302 - D_loss_real: 365.0844 - D_loss_gen: 360.1542 - DG_loss: 345.3772\n",
      "Epoch 39 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 5.1530 - D_loss_real: 364.2104 - D_loss_gen: 359.0574 - DG_loss: 344.9353\n",
      "Epoch 40 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 5.5834 - D_loss_real: 362.6978 - D_loss_gen: 357.1144 - DG_loss: 344.1043\n",
      "Epoch 41 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 6.7732 - D_loss_real: 359.4986 - D_loss_gen: 352.7254 - DG_loss: 341.7403\n",
      "Epoch 42 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 5.9775 - D_loss_real: 359.8144 - D_loss_gen: 353.8369 - DG_loss: 341.3987\n",
      "Epoch 43 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 5.6743 - D_loss_real: 361.0402 - D_loss_gen: 355.3659 - DG_loss: 341.7224\n",
      "Epoch 44 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 5.6930 - D_loss_real: 362.0399 - D_loss_gen: 356.3469 - DG_loss: 342.0443\n",
      "Epoch 45 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 5.4907 - D_loss_real: 362.3660 - D_loss_gen: 356.8753 - DG_loss: 342.3366\n",
      "Epoch 46 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 5.6463 - D_loss_real: 362.1366 - D_loss_gen: 356.4903 - DG_loss: 342.0991\n",
      "Epoch 47 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 5.7718 - D_loss_real: 361.9782 - D_loss_gen: 356.2064 - DG_loss: 342.1024\n",
      "Epoch 48 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 5.9643 - D_loss_real: 360.9224 - D_loss_gen: 354.9580 - DG_loss: 342.6900\n",
      "Epoch 49 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 6.0610 - D_loss_real: 359.4330 - D_loss_gen: 353.3720 - DG_loss: 342.0456\n",
      "Epoch 50 / 50\n",
      "936/937 [============================>.] - ETA: 0s - D_loss: 6.0547 - D_loss_real: 360.5314 - D_loss_gen: 354.4767 - DG_loss: 341.8575"
     ]
    }
   ],
   "source": [
    "g_it = 0\n",
    "\n",
    "for epoch in range(EPOCHS):\n",
    "\n",
    "    print('\\nEpoch {} / {}'.format(epoch + 1, EPOCHS))\n",
    "\n",
    "    batches_num = int(train_num_samples / BATCH_SIZE)\n",
    "    progress_bar = Progbar(target=batches_num)\n",
    "\n",
    "    epoch_D_losses_real = []\n",
    "    epoch_D_losses_gen = []\n",
    "\n",
    "    for batch_i in range(batches_num):\n",
    "\n",
    "        if len(epoch_D_losses_real) > 0:\n",
    "            progress_bar.update(\n",
    "                batch_i,\n",
    "                values=[\n",
    "                        ('D_loss', -np.mean(epoch_D_losses_real[:])  - np.mean(epoch_D_losses_gen[:])),\n",
    "                        ('D_loss_real', -np.mean(epoch_D_losses_real[:])),\n",
    "                        ('D_loss_gen', np.mean(epoch_D_losses_gen[:])),\n",
    "                        ('DG_loss', -DG_loss)\n",
    "                ]\n",
    "            )\n",
    "        else:\n",
    "            progress_bar.update(batch_i)\n",
    "            \n",
    "        # 1: train D on real+generated images\n",
    "        \n",
    "        # unfreeze D\n",
    "        D.trainable = True\n",
    "        \n",
    "        d_iterations = D_TRAIN_ITERATIONS\n",
    "        \n",
    "        # pick number of D train iterations\n",
    "        if g_it < 25 or g_it % 500 == 0:\n",
    "            d_iterations = 100\n",
    "        \n",
    "        for d_it in range(d_iterations):\n",
    "\n",
    "            # create batches of images and labels\n",
    "            real_images = X_train[batch_i * BATCH_SIZE:(batch_i + 1) * BATCH_SIZE]\n",
    "\n",
    "            # new batch of z's\n",
    "            zz = np.random.normal(0., 1., (BATCH_SIZE, Z_SIZE))\n",
    "            \n",
    "            # batch of generated images\n",
    "            generated_images = G.predict([zz])\n",
    "            \n",
    "            # clip D weights\n",
    "\n",
    "            for l in D.layers:\n",
    "                weights = l.get_weights()\n",
    "                weights = [np.clip(w, D_WEIGHTS_LOWER_BOUND, D_WEIGHTS_UPPER_BOUND) for w in weights]\n",
    "                l.set_weights(weights)\n",
    "            \n",
    "            # train D on real\n",
    "            D_loss_real = D.train_on_batch(real_images, -np.ones([BATCH_SIZE]))\n",
    "            epoch_D_losses_real.append(D_loss_real)\n",
    "            \n",
    "            # train G on generated\n",
    "            D_loss_gen = D.train_on_batch(generated_images, np.ones([BATCH_SIZE]))\n",
    "            epoch_D_losses_gen.append(D_loss_gen)\n",
    "\n",
    "        \n",
    "        # 2: train D(G) (only G is trainable)\n",
    "                \n",
    "        # freeze D\n",
    "        D.trainable = False\n",
    "        \n",
    "        DG_loss = combined.train_on_batch(\n",
    "            [np.random.normal(0., 1.,  (BATCH_SIZE, Z_SIZE))], [-np.ones(BATCH_SIZE)]\n",
    "        ) # all non-fake \n",
    "        \n",
    "        # unfreeze D\n",
    "        D.trainable = True\n",
    "        \n",
    "        if (g_it % 250 == 0):\n",
    "            save_generated_samples(g_it)\n",
    "        \n",
    "        g_it += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# save_generated_samples(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "gist": {
   "data": {
    "description": "Keras  WACGAN",
    "public": false
   },
   "id": ""
  },
  "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}