edge2nike.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf\n",
    "import scipy\n",
    "import os\n",
    "from sklearn.model_selection import train_test_split\n",
    "import time\n",
    "from IPython import display\n",
    "\n",
    "from keras.layers.convolutional import Convolution2D\n",
    "from keras.layers.convolutional import Deconvolution2D\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.layers.core import Activation\n",
    "from keras.layers.advanced_activations import LeakyReLU\n",
    "from keras.layers import Dropout\n",
    "from keras.layers import Input,merge\n",
    "from keras.models import Model\n",
    "from keras.optimizers import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Define parameters/constants\n",
    "SEED = 1337\n",
    "\n",
    "#Image Properties\n",
    "IMG_HEIGHT = 256 \n",
    "IMG_WIDTH = 256\n",
    "IMG_DEPTH = 3\n",
    "\n",
    "#Training Properties\n",
    "NUM_EPOCHS = 15\n",
    "BATCH_SIZE = 1\n",
    "\n",
    "#Convolution Properties\n",
    "KERNEL_SIZE = (4,4)\n",
    "STRIDE = (2,2)\n",
    "\n",
    "#Batch normalization mode\n",
    "MODE = 2  \n",
    "#LeakyRelu slope\n",
    "SLOPE = 0.2\n",
    "#Dropout rate\n",
    "DROPRATE = 0.5\n",
    "#Skip axis\n",
    "SKIP_AXIS = -1\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def normalize(imgs):\n",
    "    \"\"\"\n",
    "    Takes in images from [0-255] and normalizes to [-1,1]\n",
    "    \"\"\"\n",
    "    return 2*(imgs/255)-1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def unnormalize(imgs):\n",
    "    \"\"\"\n",
    "    Takes in images [-1, 1] and unnormalizes back to rgb [0-255]\n",
    "    \"\"\"\n",
    "    return (((imgs+1)/2)*255).astype('uint8')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "edge_train shape:  (750, 256, 256, 3)\n",
      "edge_test shape:  (84, 256, 256, 3)\n",
      "real_train shape:  (750, 256, 256, 3)\n",
      "real_test shape:  (84, 256, 256, 3)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1e2e62784e0>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUgAAACGCAYAAAC/gRNDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nNAzDiGAB0jAMI4IFSMMwjAg3ATBnhlHPVOxWAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1e28af29470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Load in edge images [0,255]\n",
    "edge_data = np.load('nike_edge_data.npy')\n",
    "#Normalize edge images to [-1,1]\n",
    "edge_data = normalize(edge_data)\n",
    "#Split edge images into training and test data 90/10\n",
    "edge_data,edge_data_test = train_test_split(edge_data,test_size=0.1, random_state=SEED)\n",
    "\n",
    "#Load in real color images [0,255]\n",
    "real_data = np.load('nike_shoe_data.npy')\n",
    "#Normalize real images to [-1,1]\n",
    "real_data = normalize(real_data)\n",
    "#Split real images into training and test data 90/10\n",
    "real_data,real_data_test = train_test_split(real_data,test_size=0.1,random_state=SEED)\n",
    "\n",
    "#Number of examples\n",
    "m = edge_data.shape[0]\n",
    "\n",
    "#Display training/test data shapes\n",
    "print(\"edge_train shape: \",edge_data.shape)\n",
    "print(\"edge_test shape: \",edge_data_test.shape)\n",
    "print(\"real_train shape: \",real_data.shape)\n",
    "print(\"real_test shape: \",real_data_test.shape)\n",
    "\n",
    "#Display the first edge image from training set\n",
    "IDX = 0\n",
    "plt.subplot(221)\n",
    "plt.imshow(unnormalize(edge_data[IDX]))\n",
    "plt.subplot(222)\n",
    "plt.imshow(unnormalize(real_data[IDX]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def generator_output(g_input):\n",
    "    \"\"\" \n",
    "    Creates generator network.\n",
    "    Encoder ~ C64-C128-C256-C512-C512-C512-C512-C512\n",
    "    U-Net Decoder ~ CD512-CD1024-CD1024-C1024-C1024-C512-C256-C128\n",
    "    \n",
    "    Input: generator input tensor\n",
    "    Output: generator model\n",
    "    \"\"\"\n",
    "    #Encoder\n",
    "    l1 = Convolution2D(64,*KERNEL_SIZE,border_mode='same',subsample=STRIDE)(g_input)\n",
    "    l1 = LeakyReLU(SLOPE)(l1)\n",
    "\n",
    "    l2 = Convolution2D(128,*KERNEL_SIZE,border_mode='same',subsample=STRIDE)(l1)\n",
    "    l2 = BatchNormalization(mode=MODE)(l2)\n",
    "    l2 = LeakyReLU(SLOPE)(l2)\n",
    " \n",
    "    l3 = Convolution2D(256,*KERNEL_SIZE,border_mode='same',subsample=STRIDE)(l2)\n",
    "    l3 = BatchNormalization(mode=MODE)(l3)\n",
    "    l3 = LeakyReLU(SLOPE)(l3)\n",
    "    \n",
    "    l4 = Convolution2D(512,*KERNEL_SIZE,border_mode='same',subsample=STRIDE)(l3)\n",
    "    l4 = BatchNormalization(mode=MODE)(l4)\n",
    "    l4 = LeakyReLU(SLOPE)(l4)\n",
    "    \n",
    "    l5 = Convolution2D(512,*KERNEL_SIZE,border_mode='same',subsample=STRIDE)(l4)\n",
    "    l5 = BatchNormalization(mode=MODE)(l5)\n",
    "    l5 = LeakyReLU(SLOPE)(l5)\n",
    "    \n",
    "    l6 = Convolution2D(512,*KERNEL_SIZE,border_mode='same',subsample=STRIDE)(l5)\n",
    "    l6 = BatchNormalization(mode=MODE)(l6)\n",
    "    l6 = LeakyReLU(SLOPE)(l6)\n",
    "    \n",
    "    l7 = Convolution2D(512,*KERNEL_SIZE,border_mode='same',subsample=STRIDE)(l6)\n",
    "    l7 = BatchNormalization(mode=MODE)(l7)\n",
    "    l7 = LeakyReLU(SLOPE)(l7)\n",
    "    \n",
    "    l8 = Convolution2D(512,*KERNEL_SIZE,border_mode='same',subsample=STRIDE)(l7)\n",
    "    l8 = BatchNormalization(mode=MODE)(l8)\n",
    "    l8 = LeakyReLU(SLOPE)(l8)\n",
    "\n",
    "    #U-Net Decoder\n",
    "    x7 = Deconvolution2D(512,*KERNEL_SIZE,output_shape=(None,2,2,512),border_mode='same',subsample=STRIDE)(l8)\n",
    "    x7 = BatchNormalization(mode=MODE)(x7)\n",
    "    x7 = Dropout(DROPRATE)(x7)\n",
    "    x7 = Activation('relu')(x7)\n",
    "    x7 = merge([x7,l7],mode='concat',concat_axis=SKIP_AXIS)\n",
    "\n",
    "    \n",
    "    x6 = Deconvolution2D(512,*KERNEL_SIZE,output_shape=(None,4,4,512),border_mode='same',subsample=STRIDE)(x7)\n",
    "    x6 = BatchNormalization(mode=MODE)(x6)\n",
    "    x6 = Dropout(DROPRATE)(x6)\n",
    "    x6 = Activation('relu')(x6)\n",
    "    x6 = merge([x6,l6],mode='concat',concat_axis=SKIP_AXIS)\n",
    "    \n",
    "    x5 = Deconvolution2D(512,*KERNEL_SIZE,output_shape=(None,8,8,512),border_mode='same',subsample=STRIDE)(x6)\n",
    "    x5 = BatchNormalization(mode=MODE)(x5)\n",
    "    x5 = Dropout(DROPRATE)(x5)\n",
    "    x5 = Activation('relu')(x5)\n",
    "    x5 = merge([x5,l5],mode='concat',concat_axis=SKIP_AXIS)\n",
    "    \n",
    "    x4 = Deconvolution2D(512,*KERNEL_SIZE,output_shape=(None,16,16,512),border_mode='same',subsample=STRIDE)(x5)\n",
    "    x4 = BatchNormalization(mode=MODE)(x4)\n",
    "    x4 = Dropout(DROPRATE)(x4)\n",
    "    x4 = Activation('relu')(x4)\n",
    "    x4 = merge([x4,l4],mode='concat',concat_axis=SKIP_AXIS)\n",
    "    \n",
    "    x3 = Deconvolution2D(256,*KERNEL_SIZE,output_shape=(None,32,32,256),border_mode='same',subsample=STRIDE)(x4)\n",
    "    x3 = BatchNormalization(mode=MODE)(x3)\n",
    "    x3 = Dropout(DROPRATE)(x3)\n",
    "    x3 = Activation('relu')(x3)\n",
    "    x3 = merge([x3,l3],mode='concat',concat_axis=SKIP_AXIS)\n",
    "\n",
    "    x2 = Deconvolution2D(128,*KERNEL_SIZE,output_shape=(None,64,64,128),border_mode='same',subsample=STRIDE)(x3)\n",
    "    x2 = BatchNormalization(mode=MODE)(x2)\n",
    "    x2 = Dropout(DROPRATE)(x2)\n",
    "    x2 = Activation('relu')(x2)\n",
    "    x2 = merge([x2,l2],mode='concat',concat_axis=SKIP_AXIS)\n",
    "    \n",
    "    x1 = Deconvolution2D(64,*KERNEL_SIZE,output_shape=(None,128,128,64),border_mode='same',subsample=STRIDE)(x2)\n",
    "    x1 = BatchNormalization(mode=MODE)(x1)\n",
    "    x1 = Dropout(DROPRATE)(x1)\n",
    "    x1 = Activation('relu')(x1)\n",
    "    x1 = merge([x1,l1],mode='concat',concat_axis=SKIP_AXIS)\n",
    "    \n",
    "    x0 = Deconvolution2D(IMG_DEPTH,*KERNEL_SIZE,output_shape=(None,IMG_HEIGHT,IMG_WIDTH,IMG_DEPTH),\\\n",
    "                         border_mode='same',subsample=STRIDE)(x1)\n",
    "    x0 = Activation('tanh')(x0)\n",
    "\n",
    "    return x0\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def discriminator_output(d_input):\n",
    "    \"\"\"\n",
    "    Creates discriminator network.\n",
    "    70x70 PatchGAN discriminator ~ C64-C128-C256-C512\n",
    "    Input: Edge and image concatenated: (IMG_HEIGHT,IMG_WIDTH,2*IMG_DEPTH)\n",
    "    \"\"\"\n",
    "    #\n",
    "    l1 = Convolution2D(64,*KERNEL_SIZE,border_mode='same',subsample=STRIDE)(d_input)\n",
    "    l1 = LeakyReLU(SLOPE)(l1)\n",
    "    \n",
    "    l2 = Convolution2D(128,*KERNEL_SIZE,border_mode='same',subsample=STRIDE)(l1)\n",
    "    l2 = BatchNormalization(mode=MODE)(l2)\n",
    "    l2 = LeakyReLU(SLOPE)(l2)\n",
    " \n",
    "    l3 = Convolution2D(256,*KERNEL_SIZE,border_mode='same',subsample=STRIDE)(l2)\n",
    "    l3 = BatchNormalization(mode=MODE)(l3)\n",
    "    l3 = LeakyReLU(SLOPE)(l3)\n",
    "    \n",
    "    l4 = Convolution2D(512,*KERNEL_SIZE,border_mode='same')(l3)\n",
    "    l4 = BatchNormalization(mode=MODE)(l4)\n",
    "    l4 = LeakyReLU(SLOPE)(l4)\n",
    "    \n",
    "    l5 = Convolution2D(1,1,1,border_mode='same')(l4)\n",
    "#     l5 = Activation('sigmoid')(l5)\n",
    "    return l5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Custom Loss Functions\n",
    "#L1 loss\n",
    "def l1_loss(y_true, y_pred):\n",
    "    delta = 1e-5\n",
    "    return tf.multiply(delta, tf.reduce_sum(tf.abs(y_pred - y_true)))\n",
    "\n",
    "#PatchGAN loss\n",
    "def patch_loss(y_true, y_pred):\n",
    "    loss = tf.nn.sigmoid_cross_entropy_with_logits(labels=y_true,logits=y_pred)\n",
    "    return 2*tf.reduce_mean(loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "input_1 (InputLayer)             (None, 256, 256, 3)   0                                            \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_1 (Convolution2D)  (None, 128, 128, 64)  3136        input_1[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_1 (LeakyReLU)          (None, 128, 128, 64)  0           convolution2d_1[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_2 (Convolution2D)  (None, 64, 64, 128)   131200      leakyrelu_1[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_1 (BatchNorma (None, 64, 64, 128)   512         convolution2d_2[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_2 (LeakyReLU)          (None, 64, 64, 128)   0           batchnormalization_1[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_3 (Convolution2D)  (None, 32, 32, 256)   524544      leakyrelu_2[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_2 (BatchNorma (None, 32, 32, 256)   1024        convolution2d_3[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_3 (LeakyReLU)          (None, 32, 32, 256)   0           batchnormalization_2[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_4 (Convolution2D)  (None, 16, 16, 512)   2097664     leakyrelu_3[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_3 (BatchNorma (None, 16, 16, 512)   2048        convolution2d_4[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_4 (LeakyReLU)          (None, 16, 16, 512)   0           batchnormalization_3[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_5 (Convolution2D)  (None, 8, 8, 512)     4194816     leakyrelu_4[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_4 (BatchNorma (None, 8, 8, 512)     2048        convolution2d_5[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_5 (LeakyReLU)          (None, 8, 8, 512)     0           batchnormalization_4[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_6 (Convolution2D)  (None, 4, 4, 512)     4194816     leakyrelu_5[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_5 (BatchNorma (None, 4, 4, 512)     2048        convolution2d_6[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_6 (LeakyReLU)          (None, 4, 4, 512)     0           batchnormalization_5[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_7 (Convolution2D)  (None, 2, 2, 512)     4194816     leakyrelu_6[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_6 (BatchNorma (None, 2, 2, 512)     2048        convolution2d_7[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_7 (LeakyReLU)          (None, 2, 2, 512)     0           batchnormalization_6[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_8 (Convolution2D)  (None, 1, 1, 512)     4194816     leakyrelu_7[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_7 (BatchNorma (None, 1, 1, 512)     2048        convolution2d_8[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_8 (LeakyReLU)          (None, 1, 1, 512)     0           batchnormalization_7[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "deconvolution2d_1 (Deconvolution (None, 2, 2, 512)     4194816     leakyrelu_8[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_8 (BatchNorma (None, 2, 2, 512)     2048        deconvolution2d_1[0][0]          \n",
      "____________________________________________________________________________________________________\n",
      "dropout_1 (Dropout)              (None, 2, 2, 512)     0           batchnormalization_8[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "activation_1 (Activation)        (None, 2, 2, 512)     0           dropout_1[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "merge_1 (Merge)                  (None, 2, 2, 1024)    0           activation_1[0][0]               \n",
      "                                                                   leakyrelu_7[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "deconvolution2d_2 (Deconvolution (None, 4, 4, 512)     8389120     merge_1[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_9 (BatchNorma (None, 4, 4, 512)     2048        deconvolution2d_2[0][0]          \n",
      "____________________________________________________________________________________________________\n",
      "dropout_2 (Dropout)              (None, 4, 4, 512)     0           batchnormalization_9[0][0]       \n",
      "____________________________________________________________________________________________________\n",
      "activation_2 (Activation)        (None, 4, 4, 512)     0           dropout_2[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "merge_2 (Merge)                  (None, 4, 4, 1024)    0           activation_2[0][0]               \n",
      "                                                                   leakyrelu_6[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "deconvolution2d_3 (Deconvolution (None, 8, 8, 512)     8389120     merge_2[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_10 (BatchNorm (None, 8, 8, 512)     2048        deconvolution2d_3[0][0]          \n",
      "____________________________________________________________________________________________________\n",
      "dropout_3 (Dropout)              (None, 8, 8, 512)     0           batchnormalization_10[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "activation_3 (Activation)        (None, 8, 8, 512)     0           dropout_3[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "merge_3 (Merge)                  (None, 8, 8, 1024)    0           activation_3[0][0]               \n",
      "                                                                   leakyrelu_5[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "deconvolution2d_4 (Deconvolution (None, 16, 16, 512)   8389120     merge_3[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_11 (BatchNorm (None, 16, 16, 512)   2048        deconvolution2d_4[0][0]          \n",
      "____________________________________________________________________________________________________\n",
      "dropout_4 (Dropout)              (None, 16, 16, 512)   0           batchnormalization_11[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "activation_4 (Activation)        (None, 16, 16, 512)   0           dropout_4[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "merge_4 (Merge)                  (None, 16, 16, 1024)  0           activation_4[0][0]               \n",
      "                                                                   leakyrelu_4[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "deconvolution2d_5 (Deconvolution (None, 32, 32, 256)   4194560     merge_4[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_12 (BatchNorm (None, 32, 32, 256)   1024        deconvolution2d_5[0][0]          \n",
      "____________________________________________________________________________________________________\n",
      "dropout_5 (Dropout)              (None, 32, 32, 256)   0           batchnormalization_12[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "activation_5 (Activation)        (None, 32, 32, 256)   0           dropout_5[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "merge_5 (Merge)                  (None, 32, 32, 512)   0           activation_5[0][0]               \n",
      "                                                                   leakyrelu_3[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "deconvolution2d_6 (Deconvolution (None, 64, 64, 128)   1048704     merge_5[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_13 (BatchNorm (None, 64, 64, 128)   512         deconvolution2d_6[0][0]          \n",
      "____________________________________________________________________________________________________\n",
      "dropout_6 (Dropout)              (None, 64, 64, 128)   0           batchnormalization_13[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "activation_6 (Activation)        (None, 64, 64, 128)   0           dropout_6[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "merge_6 (Merge)                  (None, 64, 64, 256)   0           activation_6[0][0]               \n",
      "                                                                   leakyrelu_2[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "deconvolution2d_7 (Deconvolution (None, 128, 128, 64)  262208      merge_6[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_14 (BatchNorm (None, 128, 128, 64)  256         deconvolution2d_7[0][0]          \n",
      "____________________________________________________________________________________________________\n",
      "dropout_7 (Dropout)              (None, 128, 128, 64)  0           batchnormalization_14[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "activation_7 (Activation)        (None, 128, 128, 64)  0           dropout_7[0][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "merge_7 (Merge)                  (None, 128, 128, 128) 0           activation_7[0][0]               \n",
      "                                                                   leakyrelu_1[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "deconvolution2d_8 (Deconvolution (None, 256, 256, 3)   6147        merge_7[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "activation_8 (Activation)        (None, 256, 256, 3)   0           deconvolution2d_8[0][0]          \n",
      "====================================================================================================\n",
      "Total params: 54,431,363\n",
      "Trainable params: 54,420,483\n",
      "Non-trainable params: 10,880\n",
      "____________________________________________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "#Build generator network\n",
    "g_input = Input(shape=(IMG_HEIGHT,IMG_WIDTH,IMG_DEPTH))\n",
    "g = generator_output(g_input)\n",
    "g_model = Model(input=g_input,output=g,name='generator')\n",
    "adam = Adam(lr=0.0002, beta_1=0.5)\n",
    "g_model.compile(optimizer = adam, loss='binary_crossentropy') #loss doesn't matter for generator\n",
    "print(g_model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "input_2 (InputLayer)             (None, 256, 256, 6)   0                                            \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_9 (Convolution2D)  (None, 128, 128, 64)  6208        input_2[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_9 (LeakyReLU)          (None, 128, 128, 64)  0           convolution2d_9[0][0]            \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_10 (Convolution2D) (None, 64, 64, 128)   131200      leakyrelu_9[0][0]                \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_15 (BatchNorm (None, 64, 64, 128)   512         convolution2d_10[0][0]           \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_10 (LeakyReLU)         (None, 64, 64, 128)   0           batchnormalization_15[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_11 (Convolution2D) (None, 32, 32, 256)   524544      leakyrelu_10[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_16 (BatchNorm (None, 32, 32, 256)   1024        convolution2d_11[0][0]           \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_11 (LeakyReLU)         (None, 32, 32, 256)   0           batchnormalization_16[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_12 (Convolution2D) (None, 32, 32, 512)   2097664     leakyrelu_11[0][0]               \n",
      "____________________________________________________________________________________________________\n",
      "batchnormalization_17 (BatchNorm (None, 32, 32, 512)   2048        convolution2d_12[0][0]           \n",
      "____________________________________________________________________________________________________\n",
      "leakyrelu_12 (LeakyReLU)         (None, 32, 32, 512)   0           batchnormalization_17[0][0]      \n",
      "____________________________________________________________________________________________________\n",
      "convolution2d_13 (Convolution2D) (None, 32, 32, 1)     513         leakyrelu_12[0][0]               \n",
      "====================================================================================================\n",
      "Total params: 2,763,713\n",
      "Trainable params: 2,761,921\n",
      "Non-trainable params: 1,792\n",
      "____________________________________________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "#Build discriminator network\n",
    "d_input = Input(shape=(IMG_HEIGHT,IMG_WIDTH,2*IMG_DEPTH))\n",
    "d = discriminator_output(d_input)\n",
    "d_model = Model(input=d_input,output=d,name='discriminator')\n",
    "adam = Adam(lr=0.0002, beta_1=0.5)\n",
    "d_model.compile(optimizer = adam, loss=patch_loss)\n",
    "print(d_model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "____________________________________________________________________________________________________\n",
      "Layer (type)                     Output Shape          Param #     Connected to                     \n",
      "====================================================================================================\n",
      "input_1 (InputLayer)             (None, 256, 256, 3)   0                                            \n",
      "____________________________________________________________________________________________________\n",
      "generator (Model)                (None, 256, 256, 3)   54431363    input_1[0][0]                    \n",
      "____________________________________________________________________________________________________\n",
      "merge_8 (Merge)                  (None, 256, 256, 6)   0           input_1[0][0]                    \n",
      "                                                                   generator[1][0]                  \n",
      "____________________________________________________________________________________________________\n",
      "discriminator (Model)            (None, 32, 32, 1)     2763713     merge_8[0][0]                    \n",
      "====================================================================================================\n",
      "Total params: 57,195,076\n",
      "Trainable params: 57,182,404\n",
      "Non-trainable params: 12,672\n",
      "____________________________________________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "#Build gan\n",
    "d_model.trainable = False\n",
    "generated_img = g_model(g_input)\n",
    "d_input_pair = merge([g_input,generated_img],mode='concat',concat_axis=3)\n",
    "\n",
    "gan_output = d_model(d_input_pair)\n",
    "gan = Model(input=g_input,output=[generated_img,gan_output],name='gan')\n",
    "gan.compile(optimizer = adam, loss=[l1_loss,patch_loss])\n",
    "print(gan.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Load in weights if not training from scratch\n",
    "g_model.load_weights('params_generator_weights_11.hdf5',True)\n",
    "d_model.load_weights('params_discriminator_weights_11.hdf5',True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAACQCAYAAAAYwZBzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1e2e7a1ce10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Initial edge input to generator\n",
    "IDX = 0\n",
    "edge_data_batch = edge_data[IDX:BATCH_SIZE+IDX,:,:,:]\n",
    "gg = g_model.predict(edge_data_batch)\n",
    "\n",
    "comb = np.concatenate([edge_data_batch[0],real_data[IDX:BATCH_SIZE+IDX,:,:,:][0]],axis=1)\n",
    "comb = np.concatenate([comb,gg[0]],axis=1)\n",
    "plt.imshow(unnormalize(comb))\n",
    "scipy.misc.imsave('nike_initial.jpg',unnormalize(comb))\n",
    "# plt.subplot(221)\n",
    "# plt.imshow(unnormalize(edge_data_batch[0]))\n",
    "# plt.subplot(222)\n",
    "# plt.imshow(unnormalize(gg[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#Try to train discriminator - one epoch\n",
    "for i in range(m):\n",
    "    edge_data_batch = edge_data[i:BATCH_SIZE+i,:,:,:]\n",
    "    real_data_batch = real_data[i:BATCH_SIZE+i,:,:,:]\n",
    "    fake_data_batch = g_model.predict(edge_data_batch)\n",
    "#     print(edge_data_batch.shape)\n",
    "#     print(real_data_batch.shape)\n",
    "\n",
    "    #Does it matter which one goes first? (edge,img), (img,edge)?\n",
    "    real_edge_pair_batch = np.concatenate((edge_data_batch,real_data_batch),axis=-1)\n",
    "    fake_edge_pair_batch = np.concatenate((edge_data_batch,fake_data_batch),axis=-1)\n",
    "    # img_pair_batch = np.concatenate((real_edge_pair_batch,fake_edge_pair_batch),axis=0)\n",
    "#     print(real_edge_pair_batch.shape)\n",
    "#     print(fake_edge_pair_batch.shape)\n",
    "\n",
    "    #Training discriminator\n",
    "    y_real = np.array([[[[1]] * 32] *32]* BATCH_SIZE)\n",
    "    d_model.train_on_batch(real_edge_pair_batch,y_real)\n",
    "\n",
    "    y_fake = np.array([[[[0]] * 32] *32]* BATCH_SIZE)\n",
    "    d_loss = d_model.train_on_batch(fake_edge_pair_batch,y_fake)\n",
    "    print(\"Step %d | D Loss: %f\"% (i,d_loss))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "lossD = []\n",
    "lossG = []\n",
    "#Loop through all epochs\n",
    "for e in range(NUM_EPOCHS):\n",
    "    for i in range(m):\n",
    "#         print(i)\n",
    "        #TODO: right now just hardcode and use batch size of 1\n",
    "        #Get edge and real image batches\n",
    "        edge_data_batch = edge_data[i:BATCH_SIZE+i,:,:,:]\n",
    "        real_data_batch = real_data[i:BATCH_SIZE+i,:,:,:]\n",
    "        \n",
    "        #Train discriminator\n",
    "        #Generate image using edges\n",
    "        fake_data_batch = g_model.predict(edge_data_batch)\n",
    "        \n",
    "        #Create (edge,real) data pair\n",
    "        real_edge_pair_batch = np.concatenate((edge_data_batch,real_data_batch),axis=-1)\n",
    "        y_real = np.array([[[[1]] * 32] *32]* BATCH_SIZE)\n",
    "        d_loss = d_model.train_on_batch(real_edge_pair_batch,y_real)\n",
    "#         print(\"Dloss \",d_loss)\n",
    "        \n",
    "        fake_edge_pair_batch = np.concatenate((edge_data_batch,fake_data_batch),axis=-1) \n",
    "        y_fake = np.array([[[[0]] * 32] *32]* BATCH_SIZE)\n",
    "        d_model.train_on_batch(fake_edge_pair_batch,y_fake)\n",
    "        \n",
    "        \n",
    "        #Train generator\n",
    "        y_real = np.array([[[[1]] * 32] *32]* BATCH_SIZE)\n",
    "#         g_loss = gan.train_on_batch(edge_data_batch, [edge_data_batch,y_real])\n",
    "        g_loss = gan.train_on_batch(edge_data_batch, [real_data_batch,y_real])\n",
    "    \n",
    "#         lossD.append(d_loss)\n",
    "#         lossG.append(g_loss[0])\n",
    "        print(\"Epoch %d Step %d | D Loss: %f G Loss (Total): %f\"% (e,i,d_loss,g_loss[0]))\n",
    "        if(i%5): #TODO: this is actually updating every step lolgg -- maybe save pics to folder instead\n",
    "            IDX = 3\n",
    "            edge_data_batch = edge_data_test[IDX:BATCH_SIZE+IDX,:,:,:]\n",
    "            real_data_batch = real_data_test[IDX:BATCH_SIZE+IDX,:,:,:]\n",
    "            fake_data_batch = g_model.predict(edge_data_batch)\n",
    "            \n",
    "            comb = np.concatenate([edge_data_batch[0],real_data_batch[0]],axis=1)\n",
    "            comb = np.concatenate([comb,fake_data_batch[0]],axis=1)\n",
    "            plt.imshow(unnormalize(comb))\n",
    "#             plt.subplot(221)\n",
    "#             plt.imshow(unnormalize(fake_data_batch[0]))\n",
    "#             plt.subplot(222)\n",
    "#             plt.imshow(unnormalize(real_data_batch[0]))\n",
    "            display.display(plt.gcf())\n",
    "            display.clear_output(wait=True)\n",
    "#             time.sleep(1)#TODO: does this need to be 2 seconds? can it be shorter?\n",
    "        \n",
    "    # save weights every epoch\n",
    "    g_model.save_weights('params_generator_weights_11.hdf5', True)\n",
    "    d_model.save_weights('params_discriminator_weights_11.hdf5', True)       \n",
    "    \n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# g_model.load_weights('params_generator_weights_6.hdf5',True)\n",
    "IDX = 0 #Test image IDX (ie. 3)\n",
    "edge_data_batch = edge_data_test[IDX:BATCH_SIZE+IDX,:,:,:]\n",
    "real_data_batch = real_data_test[IDX:BATCH_SIZE+IDX,:,:,:]\n",
    "fake_data_batch = g_model.predict(edge_data_batch)\n",
    "# fake_data_batch = ((fake_data_batch+1)/2)*255\n",
    "plt.imshow(unnormalize(fake_data_batch[0]))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# g_model.load_weights('params_generator_weights_6.hdf5',True)\n",
    "IDX = 50 #Test image IDX (ie. 3)\n",
    "edge_data_batch = edge_data_test[IDX:BATCH_SIZE+IDX,:,:,:]\n",
    "real_data_batch = real_data_test[IDX:BATCH_SIZE+IDX,:,:,:]\n",
    "fake_data_batch = g_model.predict(edge_data_batch)\n",
    "# fake_data_batch = ((fake_data_batch+1)/2)*255\n",
    "plt.imshow(unnormalize(fake_data_batch[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# g_model.load_weights('params_generator_weights_6.hdf5',True)\n",
    "#FIRE 2\n",
    "IDX =29#Test image IDX (ie. 3)\n",
    "edge_data_batch = edge_data_test[IDX:BATCH_SIZE+IDX,:,:,:]\n",
    "real_data_batch = real_data_test[IDX:BATCH_SIZE+IDX,:,:,:]\n",
    "fake_data_batch = g_model.predict(edge_data_batch)\n",
    "# fake_data_batch = ((fake_data_batch+1)/2)*255\n",
    "plt.imshow(unnormalize(fake_data_batch[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "plt.imshow(unnormalize(real_data_batch[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "plt.plot(lossD)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "plt.plot(lossG)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# save weights every epoch\n",
    "g_model.save_weights('params_generator_weights_nikelogogud.hdf5', True)\n",
    "d_model.save_weights('params_discriminator_weights_nikelogogud.hdf5', True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#Output all test image results\n",
    "#Load in weights \n",
    "g_model.load_weights('params_generator_weights_11.hdf5',True)\n",
    "d_model.load_weights('params_discriminator_weights_11.hdf5',True)\n",
    "out_dir = 'out_112/'\n",
    "\n",
    "for i in range(edge_data_test.shape[0]):\n",
    "    print(i)\n",
    "    edge_data_batch = edge_data_test[i:BATCH_SIZE+i,:,:,:]\n",
    "    real_data_batch = real_data_test[i:BATCH_SIZE+i,:,:,:]\n",
    "    fake_data_batch = g_model.predict(edge_data_batch)\n",
    "\n",
    "    comb = np.concatenate([edge_data_batch[0],real_data_batch[0]],axis=1)\n",
    "    comb = np.concatenate([comb,fake_data_batch[0]],axis=1)\n",
    "    scipy.misc.imsave(out_dir+str(i)+'.jpg',unnormalize(comb))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}