CNN Testing.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['abbey', 'airport_terminal', 'amphitheater', 'amusement_park', 'aquarium', 'aqueduct', 'art_gallery', 'assembly_line', 'auditorium']\n",
      "Loading Topic: abbey\n",
      "Loading Topic: airport_terminal\n",
      "Loading Topic: amphitheater\n",
      "Loading Topic: amusement_park\n",
      "Loading Topic: aquarium\n",
      "Loading Topic: aqueduct\n",
      "Loading Topic: art_gallery\n",
      "Loading Topic: assembly_line\n",
      "Loading Topic: auditorium\n",
      "(9000, 128, 128, 3)\n",
      "['abbey', 'airport_terminal', 'amphitheater', 'amusement_park', 'aquarium', 'aqueduct', 'art_gallery', 'assembly_line', 'auditorium']\n",
      "Loading Topic: abbey\n",
      "Loading Topic: airport_terminal\n",
      "Loading Topic: amphitheater\n",
      "Loading Topic: amusement_park\n",
      "Loading Topic: aquarium\n",
      "Loading Topic: aqueduct\n",
      "Loading Topic: art_gallery\n",
      "Loading Topic: assembly_line\n",
      "Loading Topic: auditorium\n",
      "(9000, 128, 128, 3)\n",
      "['abbey', 'airport_terminal', 'amphitheater', 'amusement_park', 'aquarium', 'aqueduct', 'art_gallery', 'assembly_line', 'auditorium']\n",
      "Loading Topic: abbey\n",
      "Loading Topic: airport_terminal\n",
      "Loading Topic: amphitheater\n",
      "Loading Topic: amusement_park\n",
      "Loading Topic: aquarium\n",
      "Loading Topic: aqueduct\n",
      "Loading Topic: art_gallery\n",
      "Loading Topic: assembly_line\n",
      "Loading Topic: auditorium\n",
      "(9000, 128, 128, 3)\n",
      "['abbey', 'airport_terminal', 'amphitheater', 'amusement_park', 'aquarium', 'aqueduct', 'art_gallery', 'assembly_line', 'auditorium']\n",
      "Loading Topic: abbey\n",
      "Loading Topic: airport_terminal\n",
      "Loading Topic: amphitheater\n",
      "Loading Topic: amusement_park\n",
      "Loading Topic: aquarium\n",
      "Loading Topic: aqueduct\n",
      "Loading Topic: art_gallery\n",
      "Loading Topic: assembly_line\n",
      "Loading Topic: auditorium\n",
      "(9000, 128, 128, 3)\n",
      "['abbey', 'airport_terminal', 'amphitheater', 'amusement_park', 'aquarium', 'aqueduct', 'art_gallery', 'assembly_line', 'auditorium']\n",
      "Loading Topic: abbey\n",
      "Loading Topic: airport_terminal\n",
      "Loading Topic: amphitheater\n",
      "Loading Topic: amusement_park\n",
      "Loading Topic: aquarium\n",
      "Loading Topic: aqueduct\n",
      "Loading Topic: art_gallery\n",
      "Loading Topic: assembly_line\n",
      "Loading Topic: auditorium\n",
      "(9000, 128, 128, 3)\n",
      "['abbey', 'airport_terminal', 'amphitheater', 'amusement_park', 'aquarium', 'aqueduct', 'art_gallery', 'assembly_line', 'auditorium']\n",
      "Loading Topic: abbey\n",
      "Loading Topic: airport_terminal\n",
      "Loading Topic: amphitheater\n",
      "Loading Topic: amusement_park\n",
      "Loading Topic: aquarium\n",
      "Loading Topic: aqueduct\n",
      "Loading Topic: art_gallery\n",
      "Loading Topic: assembly_line\n",
      "Loading Topic: auditorium\n",
      "(9000, 128, 128, 3)\n",
      "['abbey', 'airport_terminal', 'amphitheater', 'amusement_park', 'aquarium', 'aqueduct', 'art_gallery', 'assembly_line', 'auditorium']\n",
      "Loading Topic: abbey\n",
      "Loading Topic: airport_terminal\n",
      "Loading Topic: amphitheater\n",
      "Loading Topic: amusement_park\n",
      "Loading Topic: aquarium\n",
      "Loading Topic: aqueduct\n",
      "Loading Topic: art_gallery\n",
      "Loading Topic: assembly_line\n",
      "Loading Topic: auditorium\n",
      "(9000, 128, 128, 3)\n"
     ]
    }
   ],
   "source": [
    "from loadFilteredData import *\n",
    "#loading all data\n",
    "origImg = loadAllTopicData('original')\n",
    "gingham = loadAllTopicData('gingham')\n",
    "clarendon = loadAllTopicData('clarendon')\n",
    "juno = loadAllTopicData('juno')\n",
    "lark = loadAllTopicData('lark')\n",
    "gotham = loadAllTopicData('gotham')\n",
    "reyes = loadAllTopicData('reyes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9000\n",
      "9000\n",
      "9000\n",
      "9000\n",
      "9000\n",
      "9000\n",
      "9000\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-2-c884ea6c9b3a>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m     53\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     54\u001b[0m     \u001b[1;32mreturn\u001b[0m \u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mXtest\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mytest\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mXval\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0myval\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 55\u001b[1;33m \u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mXtest\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mytest\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mxval\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0myval\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mcreateDataPlaces\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0morigImg\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mclarendon\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mgingham\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mjuno\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mlark\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mgotham\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mreyes\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m.9\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m.005\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     56\u001b[0m \u001b[1;31m# X, y, Xtest, ytest, Xval, yval = createDataPlaces([origImg, clarendon], .9, .005)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m<ipython-input-2-c884ea6c9b3a>\u001b[0m in \u001b[0;36mcreateDataPlaces\u001b[1;34m(images, trainPercentage, validationPercentage)\u001b[0m\n\u001b[0;32m     48\u001b[0m     \u001b[0myval\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mvstack\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mvalVectors\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     49\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 50\u001b[1;33m     \u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mshuffle\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mrandom_state\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     51\u001b[0m     \u001b[0mXtest\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mytest\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mshuffle\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mXtest\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mytest\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mrandom_state\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     52\u001b[0m     \u001b[0mXval\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0myval\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mshuffle\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mXval\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0myval\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mrandom_state\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\sklearn\\utils\\__init__.py\u001b[0m in \u001b[0;36mshuffle\u001b[1;34m(*arrays, **options)\u001b[0m\n\u001b[0;32m    370\u001b[0m     \"\"\"\n\u001b[0;32m    371\u001b[0m     \u001b[0moptions\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'replace'\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;32mFalse\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 372\u001b[1;33m     \u001b[1;32mreturn\u001b[0m \u001b[0mresample\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m*\u001b[0m\u001b[0marrays\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m**\u001b[0m\u001b[0moptions\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    373\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    374\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\sklearn\\utils\\__init__.py\u001b[0m in \u001b[0;36mresample\u001b[1;34m(*arrays, **options)\u001b[0m\n\u001b[0;32m    294\u001b[0m     \u001b[1;31m# convert sparse matrices to CSR for row-based indexing\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    295\u001b[0m     \u001b[0marrays\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m[\u001b[0m\u001b[0ma\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtocsr\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mif\u001b[0m \u001b[0missparse\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0ma\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32melse\u001b[0m \u001b[0ma\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0ma\u001b[0m \u001b[1;32min\u001b[0m \u001b[0marrays\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 296\u001b[1;33m     \u001b[0mresampled_arrays\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m[\u001b[0m\u001b[0msafe_indexing\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0ma\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mindices\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0ma\u001b[0m \u001b[1;32min\u001b[0m \u001b[0marrays\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    297\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mresampled_arrays\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m==\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    298\u001b[0m         \u001b[1;31m# syntactic sugar for the unit argument case\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\sklearn\\utils\\__init__.py\u001b[0m in \u001b[0;36m<listcomp>\u001b[1;34m(.0)\u001b[0m\n\u001b[0;32m    294\u001b[0m     \u001b[1;31m# convert sparse matrices to CSR for row-based indexing\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    295\u001b[0m     \u001b[0marrays\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m[\u001b[0m\u001b[0ma\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtocsr\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mif\u001b[0m \u001b[0missparse\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0ma\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32melse\u001b[0m \u001b[0ma\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0ma\u001b[0m \u001b[1;32min\u001b[0m \u001b[0marrays\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 296\u001b[1;33m     \u001b[0mresampled_arrays\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m[\u001b[0m\u001b[0msafe_indexing\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0ma\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mindices\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0ma\u001b[0m \u001b[1;32min\u001b[0m \u001b[0marrays\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    297\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mresampled_arrays\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m==\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    298\u001b[0m         \u001b[1;31m# syntactic sugar for the unit argument case\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\sklearn\\utils\\__init__.py\u001b[0m in \u001b[0;36msafe_indexing\u001b[1;34m(X, indices)\u001b[0m\n\u001b[0;32m    183\u001b[0m                                    indices.dtype.kind == 'i'):\n\u001b[0;32m    184\u001b[0m             \u001b[1;31m# This is often substantially faster than X[indices]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 185\u001b[1;33m             \u001b[1;32mreturn\u001b[0m \u001b[0mX\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mtake\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mindices\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    186\u001b[0m         \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    187\u001b[0m             \u001b[1;32mreturn\u001b[0m \u001b[0mX\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mindices\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "def createDataPlaces(images, trainPercentage, validationPercentage):\n",
    "    categories = len(images)\n",
    "\n",
    "    imgList = []\n",
    "    vectors = []\n",
    "\n",
    "    testImgList = []\n",
    "    testVectors = []\n",
    "\n",
    "    valImgList = []\n",
    "    valVectors = []\n",
    "    \n",
    "\n",
    "    #data for original image\n",
    "    for c in range(categories):\n",
    "        numImages = images[c].shape[0]\n",
    "        print(numImages)\n",
    "        numTrain = int(numImages * trainPercentage)\n",
    "        numVal = int(numTrain * validationPercentage)\n",
    "        \n",
    "        imgList.append(images[c][:numTrain - numVal])\n",
    "        \n",
    "        valImgList.append(images[c][numTrain - numVal:numTrain])\n",
    "\n",
    "        featureVector = np.zeros((numTrain - numVal, categories))\n",
    "        featureVector[:, c] = 1\n",
    "        vectors.append(featureVector)\n",
    "        \n",
    "        featureVector = np.zeros((numVal, categories))\n",
    "        featureVector[:, c] = 1\n",
    "        valVectors.append(featureVector)\n",
    "    \n",
    "        #testing data\n",
    "        testImgList.append(images[c][numTrain:])\n",
    "\n",
    "        featureVector = np.zeros((numImages - numTrain, categories))\n",
    "        featureVector[:, c] = 1\n",
    "        testVectors.append(featureVector)\n",
    "\n",
    "\n",
    "    X = np.vstack(imgList)\n",
    "    y = np.vstack(vectors)\n",
    "\n",
    "    Xtest = np.vstack(testImgList)\n",
    "    ytest = np.vstack(testVectors)\n",
    "    \n",
    "    Xval = np.vstack(valImgList)\n",
    "    yval = np.vstack(valVectors)\n",
    "\n",
    "    X, y = shuffle(X, y, random_state=0)\n",
    "    Xtest, ytest = shuffle(Xtest, ytest, random_state=0)\n",
    "    Xval, yval = shuffle(Xval, yval, random_state=0)\n",
    "    \n",
    "    return X, y, Xtest, ytest, Xval, yval\n",
    "X, y, Xtest, ytest, xval, yval = createDataPlaces([origImg, clarendon, gingham, juno, lark, gotham, reyes], .9, .005)\n",
    "# X, y, Xtest, ytest, Xval, yval = createDataPlaces([origImg, clarendon], .9, .005)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(80, 128, 128, 3)\n",
      "(280, 7)\n"
     ]
    }
   ],
   "source": [
    "print(Xval.shape)\n",
    "print(yval.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def splitImg(X, y):\n",
    "    Xsmall = []\n",
    "    ysmall = []\n",
    "    for i in range(X.shape[0]):\n",
    "        l = 0\n",
    "        for j in range(0, X.shape[1], X.shape[1] // 4):\n",
    "            for k in range(0, X.shape[1], X.shape[1] // 4):\n",
    "                block = X[i, j : j + 32, k : k + 32]\n",
    "                Xsmall.append(block)\n",
    "                ysmall.append(y[i])\n",
    "                l += 1\n",
    "    Xsmall = np.array(Xsmall)\n",
    "    ysmall = np.array(ysmall)\n",
    "    return Xsmall, ysmall"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "Xsmall, ysmall = splitImg(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(257920, 32, 32, 3)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Xsmall.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "XvalSmall, yvalSmall = splitImg(Xval, yval)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1280, 32, 32, 3)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "XvalSmall.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Flatten\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras.layers import LeakyReLU\n",
    "import numpy as np\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Conv2D(32, (3, 3), input_shape=(32, 32, 3), strides=2))\n",
    "model.add(LeakyReLU(alpha=0.5))\n",
    "model.add(Conv2D(32, (3, 3), strides = 2))\n",
    "model.add(LeakyReLU(alpha=0.5))\n",
    "model.add(MaxPooling2D(pool_size=3, strides =2))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(128))\n",
    "model.add(LeakyReLU(alpha=0.3))\n",
    "model.add(Dense(256))\n",
    "model.add(LeakyReLU(alpha=0.3))\n",
    "# model.add(Dropout(0.5))\n",
    "model.add(Dense(2, activation='softmax'))\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 257920 samples, validate on 1280 samples\n",
      "Epoch 1/5\n",
      "257920/257920 [==============================] - 29s 113us/step - loss: 2.5714 - acc: 0.7365 - val_loss: 0.2310 - val_acc: 0.9133\n",
      "Epoch 2/5\n",
      "257920/257920 [==============================] - 26s 101us/step - loss: 0.3389 - acc: 0.8441 - val_loss: 0.1395 - val_acc: 0.9414\n",
      "Epoch 3/5\n",
      "257920/257920 [==============================] - 26s 100us/step - loss: 0.2806 - acc: 0.8661 - val_loss: 0.0863 - val_acc: 0.9648\n",
      "Epoch 4/5\n",
      "257920/257920 [==============================] - 25s 98us/step - loss: 0.2489 - acc: 0.8781 - val_loss: 0.0651 - val_acc: 0.9672\n",
      "Epoch 5/5\n",
      "257920/257920 [==============================] - 25s 98us/step - loss: 0.2373 - acc: 0.8828 - val_loss: 0.0866 - val_acc: 0.9625\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x19edc9aecc0>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(Xsmall, ysmall, epochs=5, batch_size=128, validation_data=(XvalSmall, yvalSmall))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[14256   144]\n",
      " [ 1557 12843]]\n",
      "0.9409375\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(array([0.90153671, 0.98891199]),\n",
       " array([0.99    , 0.891875]),\n",
       " array([0.94369973, 0.93789024]),\n",
       " array([14400, 14400], dtype=int64))"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "XtestSmall, ytestSmall = splitImg(Xtest, ytest)\n",
    "\n",
    "predict = model.predict(XtestSmall)\n",
    "from sklearn.metrics import confusion_matrix\n",
    "print(confusion_matrix(ytestSmall.argmax(axis=1), predict.argmax(axis=1)))\n",
    "from sklearn.metrics import accuracy_score\n",
    "print(accuracy_score(ytestSmall.argmax(axis=1), predict.argmax(axis=1)))\n",
    "from sklearn.metrics import precision_recall_fscore_support\n",
    "precision_recall_fscore_support(ytestSmall.argmax(axis=1), predict.argmax(axis=1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Flatten\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras.layers import LeakyReLU\n",
    "import numpy as np\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Conv2D(32, (3, 3), input_shape=(32, 32, 3), strides=2))\n",
    "model.add(LeakyReLU(alpha=0.3))\n",
    "model.add(Conv2D(32, (3, 3), strides = 2))\n",
    "model.add(LeakyReLU(alpha=0.3))\n",
    "model.add(MaxPooling2D(pool_size=3, strides =2))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(128))\n",
    "model.add(LeakyReLU(alpha=0.3))\n",
    "model.add(Dense(256))\n",
    "model.add(LeakyReLU(alpha=0.3))\n",
    "# model.add(Dropout(0.5))\n",
    "model.add(Dense(2, activation='softmax'))\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 256608 samples, validate on 2592 samples\n",
      "Epoch 1/150\n",
      " 34304/256608 [===>..........................] - ETA: 30s - loss: 1.3212 - acc: 0.6842"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-87-03623e499d47>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mmodel\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mXsmall\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mysmall\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mepochs\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m150\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m128\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mvalidation_split\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\keras\\models.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, **kwargs)\u001b[0m\n\u001b[0;32m    891\u001b[0m                               \u001b[0mclass_weight\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mclass_weight\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    892\u001b[0m                               \u001b[0msample_weight\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0msample_weight\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 893\u001b[1;33m                               initial_epoch=initial_epoch)\n\u001b[0m\u001b[0;32m    894\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    895\u001b[0m     def evaluate(self, x, y, batch_size=32, verbose=1,\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\keras\\engine\\training.py\u001b[0m in \u001b[0;36mfit\u001b[1;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, **kwargs)\u001b[0m\n\u001b[0;32m   1629\u001b[0m                               \u001b[0minitial_epoch\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0minitial_epoch\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1630\u001b[0m                               \u001b[0msteps_per_epoch\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0msteps_per_epoch\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1631\u001b[1;33m                               validation_steps=validation_steps)\n\u001b[0m\u001b[0;32m   1632\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1633\u001b[0m     def evaluate(self, x=None, y=None,\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\keras\\engine\\training.py\u001b[0m in \u001b[0;36m_fit_loop\u001b[1;34m(self, f, ins, out_labels, batch_size, epochs, verbose, callbacks, val_f, val_ins, shuffle, callback_metrics, initial_epoch, steps_per_epoch, validation_steps)\u001b[0m\n\u001b[0;32m   1211\u001b[0m                     \u001b[0mbatch_logs\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;34m'size'\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mbatch_ids\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1212\u001b[0m                     \u001b[0mcallbacks\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mon_batch_begin\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mbatch_index\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mbatch_logs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1213\u001b[1;33m                     \u001b[0mouts\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mf\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mins_batch\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1214\u001b[0m                     \u001b[1;32mif\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[0misinstance\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mouts\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mlist\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1215\u001b[0m                         \u001b[0mouts\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m[\u001b[0m\u001b[0mouts\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py\u001b[0m in \u001b[0;36m__call__\u001b[1;34m(self, inputs)\u001b[0m\n\u001b[0;32m   2330\u001b[0m         updated = session.run(self.outputs + [self.updates_op],\n\u001b[0;32m   2331\u001b[0m                               \u001b[0mfeed_dict\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mfeed_dict\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 2332\u001b[1;33m                               **self.session_kwargs)\n\u001b[0m\u001b[0;32m   2333\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mupdated\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;33m:\u001b[0m\u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0moutputs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   2334\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\tensorflow\\python\\client\\session.py\u001b[0m in \u001b[0;36mrun\u001b[1;34m(self, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[0;32m    776\u001b[0m     \u001b[1;32mtry\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    777\u001b[0m       result = self._run(None, fetches, feed_dict, options_ptr,\n\u001b[1;32m--> 778\u001b[1;33m                          run_metadata_ptr)\n\u001b[0m\u001b[0;32m    779\u001b[0m       \u001b[1;32mif\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    780\u001b[0m         \u001b[0mproto_data\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mtf_session\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mTF_GetBuffer\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mrun_metadata_ptr\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\tensorflow\\python\\client\\session.py\u001b[0m in \u001b[0;36m_run\u001b[1;34m(self, handle, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[0;32m    980\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mfinal_fetches\u001b[0m \u001b[1;32mor\u001b[0m \u001b[0mfinal_targets\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    981\u001b[0m       results = self._do_run(handle, final_targets, final_fetches,\n\u001b[1;32m--> 982\u001b[1;33m                              feed_dict_string, options, run_metadata)\n\u001b[0m\u001b[0;32m    983\u001b[0m     \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    984\u001b[0m       \u001b[0mresults\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m[\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\tensorflow\\python\\client\\session.py\u001b[0m in \u001b[0;36m_do_run\u001b[1;34m(self, handle, target_list, fetch_list, feed_dict, options, run_metadata)\u001b[0m\n\u001b[0;32m   1030\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[0mhandle\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1031\u001b[0m       return self._do_call(_run_fn, self._session, feed_dict, fetch_list,\n\u001b[1;32m-> 1032\u001b[1;33m                            target_list, options, run_metadata)\n\u001b[0m\u001b[0;32m   1033\u001b[0m     \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1034\u001b[0m       return self._do_call(_prun_fn, self._session, handle, feed_dict,\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\tensorflow\\python\\client\\session.py\u001b[0m in \u001b[0;36m_do_call\u001b[1;34m(self, fn, *args)\u001b[0m\n\u001b[0;32m   1037\u001b[0m   \u001b[1;32mdef\u001b[0m \u001b[0m_do_call\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m*\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1038\u001b[0m     \u001b[1;32mtry\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1039\u001b[1;33m       \u001b[1;32mreturn\u001b[0m \u001b[0mfn\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m*\u001b[0m\u001b[0margs\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   1040\u001b[0m     \u001b[1;32mexcept\u001b[0m \u001b[0merrors\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mOpError\u001b[0m \u001b[1;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1041\u001b[0m       \u001b[0mmessage\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mcompat\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mas_text\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0me\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmessage\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Local\\conda\\conda\\envs\\magiclearning\\lib\\site-packages\\tensorflow\\python\\client\\session.py\u001b[0m in \u001b[0;36m_run_fn\u001b[1;34m(session, feed_dict, fetch_list, target_list, options, run_metadata)\u001b[0m\n\u001b[0;32m   1019\u001b[0m         return tf_session.TF_Run(session, options,\n\u001b[0;32m   1020\u001b[0m                                  \u001b[0mfeed_dict\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfetch_list\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtarget_list\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1021\u001b[1;33m                                  status, run_metadata)\n\u001b[0m\u001b[0;32m   1022\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1023\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m_prun_fn\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0msession\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mhandle\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfeed_dict\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfetch_list\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "model.fit(Xsmall, ysmall, epochs=150, batch_size=128, validation_split=0.01)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9000\n",
      "9000\n",
      "9000\n",
      "9000\n",
      "9000\n",
      "9000\n",
      "9000\n"
     ]
    }
   ],
   "source": [
    "def createImageData(images, trainPercentage):\n",
    "    categories = len(images)\n",
    "\n",
    "    imgList = []\n",
    "    vectors = []\n",
    "\n",
    "    testImgList = []\n",
    "    testVectors = []\n",
    "\n",
    "    #data for original image\n",
    "    for c in range(categories):\n",
    "        numImages = images[c].shape[0]\n",
    "        print(numImages)\n",
    "        numTrain = int(numImages * trainPercentage)\n",
    "        imgList.append(images[c][:numTrain])\n",
    "\n",
    "        featureVector = np.zeros((numTrain, categories))\n",
    "        featureVector[:, c] = 1\n",
    "#         vectors.append(featureVector)\n",
    "        vectors.append(images[0][:numTrain])\n",
    "    \n",
    "        #testing data\n",
    "        testImgList.append(images[c][numTrain:])\n",
    "\n",
    "        featureVector = np.zeros((numImages - numTrain, categories))\n",
    "        featureVector[:, c] = 1\n",
    "#         testVectors.append(featureVector)\n",
    "        testVectors.append(images[0][numTrain:])\n",
    "\n",
    "\n",
    "    X = np.vstack(imgList)\n",
    "    y = np.vstack(vectors)\n",
    "\n",
    "    Xtest = np.vstack(testImgList)\n",
    "    ytest = np.vstack(testVectors)\n",
    "\n",
    "    X, y = shuffle(X, y, random_state=0)\n",
    "    Xtest, ytest = shuffle(Xtest, ytest, random_state=0)\n",
    "    return X, y, Xtest, ytest\n",
    "X, y, Xtest, ytest = createImageData([origImg, clarendon, gingham, juno, lark, gotham, reyes], .9)\n",
    "# imX, imy, imXtest, imytest = createImageData([origImg, clarendon], .9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(56700, 128, 128, 3)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def divideImages(X, y):\n",
    "    imXsmall = []\n",
    "    imysmall = []\n",
    "    for i in range(X.shape[0]):\n",
    "        l = 0\n",
    "        for j in range(0, X.shape[1], X.shape[1] // 4):\n",
    "            for k in range(0, X.shape[1], X.shape[1] // 4):\n",
    "                blockX = X[i, j : j + 32, k : k + 32]\n",
    "                blocky = y[i, j : j + 32, k : k + 32]\n",
    "                imXsmall.append(blockX)\n",
    "                imysmall.append(blocky)\n",
    "                l += 1\n",
    "    imXsmall = np.array(imXsmall)\n",
    "    imysmall = np.array(imysmall)\n",
    "    return imXsmall, imysmall\n",
    "imXsmall, imysmall = divideImages(X, y)\n",
    "imXsmall, imysmall = shuffle(imXsmall, imysmall, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "279300\n",
      "[[243 244 247 ...   6  11  22]\n",
      " [249 246 244 ...  21  22  27]\n",
      " [ 11  10  10 ...  27  25  24]\n",
      " ...\n",
      " [ 30  24  18 ...  19  20  19]\n",
      " [  5   7  11 ...  21  21  20]\n",
      " [ 11  13  16 ...  21  21  20]]\n"
     ]
    }
   ],
   "source": [
    "imsmallResidual = imysmall - imXsmall\n",
    "i = 442\n",
    "plt.imshow(imXsmall[i])\n",
    "plt.show()\n",
    "plt.imshow(imysmall[i])\n",
    "plt.show()\n",
    "print(np.sum(np.abs(imXsmall[i] - imysmall[i])))\n",
    "print(imsmallResidual[i][:,:, 0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Flatten\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras.layers import LeakyReLU\n",
    "import numpy as np\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Conv2D(6, (3, 3), input_shape=(32, 32, 3), padding='same'))\n",
    "model.add(LeakyReLU(alpha=0.3))\n",
    "model.add(Conv2D(6, (3, 3), padding='same'))\n",
    "model.add(LeakyReLU(alpha=0.3))\n",
    "model.add(Dense(128))\n",
    "model.add(LeakyReLU(alpha=0.3))\n",
    "model.add(Dense(256))\n",
    "model.add(LeakyReLU(alpha=0.3))\n",
    "model.add(Conv2D(3, (3, 3), padding='same'))\n",
    "# model.add(MaxPooling2D(pool_size=3, strides =2))\n",
    "# model.add(Flatten())\n",
    "# model.add(Dense(128))\n",
    "# model.add(LeakyReLU(alpha=0.3))\n",
    "# model.add(Dense(256))\n",
    "# model.add(LeakyReLU(alpha=0.3))\n",
    "# model.add(Conv2DTranspose(32, (3, 3), strides = 2))\n",
    "# model.add(Conv2D(32, (3, 3), input_shape=(32, 32, 3), strides=2))\n",
    "# model.add(Dropout(0.5))\n",
    "#model.add(Dense(2, activation='softmax'))\n",
    "model.compile(loss='mean_squared_error', optimizer='adam', metrics=['mae'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 898128 samples, validate on 9072 samples\n",
      "Epoch 1/1\n",
      "898128/898128 [==============================] - 929s 1ms/step - loss: 417.3434 - mean_absolute_error: 15.8734 - val_loss: 456.5389 - val_mean_absolute_error: 17.5183\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x1cdb7909518>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(imXsmall, imysmall, epochs=1, batch_size=256, validation_split=0.01)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "77.29099801558596"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(np.abs(imXsmall - imysmall))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "imXtestsmall, imytestsmall = divideImages(Xtest, ytest)\n",
    "#imXtestsmall, imytestsmall = shuffle(imXtestsmall, imytestsmall, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "14.9010105"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predict = model.predict(imXtestsmall)\n",
    "np.mean(np.abs(predict - imytestsmall))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import load_model\n",
    "model.save('EndToEndNN1.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "90.91796875\n"
     ]
    }
   ],
   "source": [
    "i = 442\n",
    "plt.imshow(imXsmall[i])\n",
    "plt.show()\n",
    "plt.imshow(imysmall[i])\n",
    "plt.show()\n",
    "print(np.mean(np.abs(imXsmall[i] - imysmall[i])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "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"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(16, 32, 32, 3)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "15.515706380208334\n"
     ]
    }
   ],
   "source": [
    "k = 5500\n",
    "im = []\n",
    "for i in range(k * 4, k * 4 + 4):\n",
    "    line = np.hstack(imytestsmall[i * 4 : i * 4 + 4])\n",
    "    im.append(line)\n",
    "im = np.vstack(im)\n",
    "\n",
    "plt.imshow(im)\n",
    "plt.show()\n",
    "\n",
    "im = []\n",
    "for i in range(k * 4, k * 4 + 4):\n",
    "    line = np.hstack(imXtestsmall[i * 4 : i * 4 + 4])\n",
    "    im.append(line)\n",
    "im = np.vstack(im)\n",
    "\n",
    "plt.imshow(im)\n",
    "plt.show()\n",
    "\n",
    "\n",
    "imPredSmall = model.predict(imXtestsmall[k * 16 : k * 16 + 16])\n",
    "imPredSmall = imPredSmall.astype('int')\n",
    "print(imPredSmall.shape)\n",
    "im = []\n",
    "for i in range(4):\n",
    "    line = np.hstack(imPredSmall[i * 4 : i * 4 + 4])\n",
    "    im.append(line)\n",
    "im = np.vstack(im)\n",
    "\n",
    "plt.imshow(im)\n",
    "plt.show()\n",
    "\n",
    "print(np.mean(np.abs(imPredSmall - imytestsmall[k * 16 : k * 16 + 16])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[[ 10,   5,  12],\n",
       "         [  0,   3,  15],\n",
       "         [  0,   6,  17],\n",
       "         ...,\n",
       "         [  3,   4,   9],\n",
       "         [  0,   6,   9],\n",
       "         [  0,  10,  11]],\n",
       "\n",
       "        [[ 17,  12,  19],\n",
       "         [  3,   9,  21],\n",
       "         [  0,  13,  24],\n",
       "         ...,\n",
       "         [  0,   5,   5],\n",
       "         [  0,   8,   3],\n",
       "         [  0,  11,   4]],\n",
       "\n",
       "        [[  0,   0,   7],\n",
       "         [  0,   4,  16],\n",
       "         [  4,  26,  39],\n",
       "         ...,\n",
       "         [  1,  27,  16],\n",
       "         [  0,  30,  16],\n",
       "         [  0,  31,  15]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[  0,   4,   5],\n",
       "         [  7,  25,  27],\n",
       "         [ 61,  90,  94],\n",
       "         ...,\n",
       "         [ 20,  20,  20],\n",
       "         [  0,   0,   4],\n",
       "         [  0,   1,   6]],\n",
       "\n",
       "        [[  0,   3,   4],\n",
       "         [  7,  25,  27],\n",
       "         [ 62,  91,  95],\n",
       "         ...,\n",
       "         [ 21,  21,  21],\n",
       "         [  1,   0,   6],\n",
       "         [  0,   1,   6]],\n",
       "\n",
       "        [[  0,   3,   4],\n",
       "         [  7,  25,  27],\n",
       "         [ 62,  91,  95],\n",
       "         ...,\n",
       "         [ 23,  23,  23],\n",
       "         [  0,   1,   6],\n",
       "         [  0,   1,   8]]],\n",
       "\n",
       "\n",
       "       [[[  2,   0,  11],\n",
       "         [  5,   0,   9],\n",
       "         [  9,   0,  10],\n",
       "         ...,\n",
       "         [  2,   1,   9],\n",
       "         [  2,   1,   9],\n",
       "         [  1,   0,   8]],\n",
       "\n",
       "        [[  0,   5,   7],\n",
       "         [  0,   5,   5],\n",
       "         [  0,   4,   5],\n",
       "         ...,\n",
       "         [  0,   6,   5],\n",
       "         [  0,   6,   5],\n",
       "         [  0,   6,   5]],\n",
       "\n",
       "        [[  0,  27,  16],\n",
       "         [  0,  27,  16],\n",
       "         [  0,  28,  17],\n",
       "         ...,\n",
       "         [  0,  26,  15],\n",
       "         [  0,  26,  15],\n",
       "         [  0,  26,  15]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[  2,   0,   3],\n",
       "         [  0,   1,   6],\n",
       "         [  0,   1,   6],\n",
       "         ...,\n",
       "         [244, 155, 123],\n",
       "         [255, 205, 156],\n",
       "         [254, 207, 153]],\n",
       "\n",
       "        [[  2,   0,   5],\n",
       "         [  0,   1,   6],\n",
       "         [  0,   1,   6],\n",
       "         ...,\n",
       "         [251, 120, 100],\n",
       "         [255, 167, 139],\n",
       "         [248, 174, 145]],\n",
       "\n",
       "        [[  1,   0,   5],\n",
       "         [  0,   1,   6],\n",
       "         [  0,   1,   6],\n",
       "         ...,\n",
       "         [255,  93,  78],\n",
       "         [243, 120, 104],\n",
       "         [187,  97,  86]]],\n",
       "\n",
       "\n",
       "       [[[  5,   0,   9],\n",
       "         [  4,   0,   8],\n",
       "         [  4,   0,   8],\n",
       "         ...,\n",
       "         [  8,   2,  16],\n",
       "         [  9,   2,  18],\n",
       "         [  9,   2,  18]],\n",
       "\n",
       "        [[  0,   4,   8],\n",
       "         [  0,   4,   8],\n",
       "         [  0,   4,   8],\n",
       "         ...,\n",
       "         [  0,   5,  11],\n",
       "         [  0,   7,  12],\n",
       "         [  0,   7,  12]],\n",
       "\n",
       "        [[  0,  22,  20],\n",
       "         [  0,  22,  20],\n",
       "         [  0,  21,  19],\n",
       "         ...,\n",
       "         [  0,  25,  19],\n",
       "         [  0,  26,  20],\n",
       "         [  0,  26,  20]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[250, 195, 138],\n",
       "         [250, 204, 145],\n",
       "         [243, 217, 156],\n",
       "         ...,\n",
       "         [192, 150, 112],\n",
       "         [171, 138,  97],\n",
       "         [163, 134,  92]],\n",
       "\n",
       "        [[208, 188, 135],\n",
       "         [200, 194, 142],\n",
       "         [186, 202, 153],\n",
       "         ...,\n",
       "         [161, 125,  89],\n",
       "         [134, 111,  70],\n",
       "         [143, 126,  82]],\n",
       "\n",
       "        [[ 86, 102,  57],\n",
       "         [ 66, 103,  62],\n",
       "         [ 45, 102,  69],\n",
       "         ...,\n",
       "         [137, 114,  72],\n",
       "         [101,  98,  55],\n",
       "         [118, 127,  82]]],\n",
       "\n",
       "\n",
       "       ...,\n",
       "\n",
       "\n",
       "       [[[200, 108,  97],\n",
       "         [192,  99,  84],\n",
       "         [174,  75,  56],\n",
       "         ...,\n",
       "         [ 21,  54,  73],\n",
       "         [ 25,  57,  78],\n",
       "         [ 32,  62,  86]],\n",
       "\n",
       "        [[214, 111, 104],\n",
       "         [196,  98,  85],\n",
       "         [189,  94,  74],\n",
       "         ...,\n",
       "         [ 59,  91, 106],\n",
       "         [ 60,  89, 105],\n",
       "         [ 59,  88, 106]],\n",
       "\n",
       "        [[227, 116, 109],\n",
       "         [194,  93,  81],\n",
       "         [203, 112,  91],\n",
       "         ...,\n",
       "         [ 74, 104, 115],\n",
       "         [ 73, 103, 114],\n",
       "         [ 67,  97, 107]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[  0,   5,  11],\n",
       "         [  0,   9,  10],\n",
       "         [  0,  12,  10],\n",
       "         ...,\n",
       "         [  0,  21,  18],\n",
       "         [  0,  21,  18],\n",
       "         [  0,  21,  18]],\n",
       "\n",
       "        [[  0,   4,   2],\n",
       "         [  0,   2,   0],\n",
       "         [  0,   3,   6],\n",
       "         ...,\n",
       "         [  0,  10,   5],\n",
       "         [  0,  10,   5],\n",
       "         [  0,  10,   5]],\n",
       "\n",
       "        [[  9,   5,   0],\n",
       "         [  5,   0,   0],\n",
       "         [  6,   0,   2],\n",
       "         ...,\n",
       "         [  3,  14,  10],\n",
       "         [  3,  14,   8],\n",
       "         [  2,  15,   8]]],\n",
       "\n",
       "\n",
       "       [[[ 50,  66,  92],\n",
       "         [ 34,  52,  76],\n",
       "         [ 42,  60,  82],\n",
       "         ...,\n",
       "         [187, 102,  81],\n",
       "         [152,  90,  77],\n",
       "         [136,  87,  83]],\n",
       "\n",
       "        [[ 55,  76,  95],\n",
       "         [ 55,  78,  94],\n",
       "         [ 67,  90, 106],\n",
       "         ...,\n",
       "         [169,  83,  66],\n",
       "         [133,  65,  62],\n",
       "         [126,  66,  74]],\n",
       "\n",
       "        [[ 72,  98, 111],\n",
       "         [ 78, 106, 118],\n",
       "         [ 81, 109, 123],\n",
       "         ...,\n",
       "         [196, 119, 109],\n",
       "         [151,  86,  90],\n",
       "         [129,  62,  80]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[  0,  19,  16],\n",
       "         [  0,  20,  18],\n",
       "         [  0,  19,  20],\n",
       "         ...,\n",
       "         [  2,  16,  17],\n",
       "         [  0,  10,   8],\n",
       "         [  0,  19,  10]],\n",
       "\n",
       "        [[  0,   8,   1],\n",
       "         [  0,   9,   2],\n",
       "         [  0,   9,   6],\n",
       "         ...,\n",
       "         [  1,   3,   0],\n",
       "         [  0,   3,   0],\n",
       "         [  0,   5,   0]],\n",
       "\n",
       "        [[  2,  13,   5],\n",
       "         [  2,  13,   7],\n",
       "         [  3,  14,  10],\n",
       "         ...,\n",
       "         [  2,   2,   2],\n",
       "         [  0,   5,   1],\n",
       "         [  0,   2,   0]]],\n",
       "\n",
       "\n",
       "       [[[206, 107,  86],\n",
       "         [240, 151, 121],\n",
       "         [244, 161, 131],\n",
       "         ...,\n",
       "         [104, 103, 101],\n",
       "         [  0,  10,  11],\n",
       "         [  0,   3,   2]],\n",
       "\n",
       "        [[226, 125, 105],\n",
       "         [247, 156, 127],\n",
       "         [247, 166, 136],\n",
       "         ...,\n",
       "         [ 98,  90,  87],\n",
       "         [  0,   2,   4],\n",
       "         [  0,   4,   5]],\n",
       "\n",
       "        [[241, 136, 117],\n",
       "         [243, 154, 124],\n",
       "         [245, 166, 135],\n",
       "         ...,\n",
       "         [117, 104,  96],\n",
       "         [  0,   3,   3],\n",
       "         [  0,   4,   6]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[  0,  15,   6],\n",
       "         [  0,  17,  13],\n",
       "         [  0,   9,   8],\n",
       "         ...,\n",
       "         [  1,  25,  37],\n",
       "         [  0,   7,  24],\n",
       "         [  0,   0,  19]],\n",
       "\n",
       "        [[  0,   2,   2],\n",
       "         [  0,   1,   0],\n",
       "         [  0,   1,   0],\n",
       "         ...,\n",
       "         [  0,   2,  12],\n",
       "         [  0,   0,  11],\n",
       "         [  0,   4,  13]],\n",
       "\n",
       "        [[  4,   2,  13],\n",
       "         [  0,   2,   5],\n",
       "         [  0,   2,   3],\n",
       "         ...,\n",
       "         [  9,   0,   0],\n",
       "         [  6,   0,   0],\n",
       "         [  5,   6,   1]]]], dtype=uint8)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "imXtestsmall[k * 16 : k * 16 + 16]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 89  35  11]\n",
      " [ 89  35  11]\n",
      " [ 89  35  11]\n",
      " [ 90  36  12]\n",
      " [ 90  36  12]\n",
      " [ 90  36  10]\n",
      " [ 91  37  11]\n",
      " [ 91  37  11]\n",
      " [ 91  37   9]\n",
      " [ 91  37   9]\n",
      " [ 92  38  10]\n",
      " [ 93  39  11]\n",
      " [ 95  42  11]\n",
      " [ 96  43  12]\n",
      " [ 97  44  13]\n",
      " [ 97  44  13]\n",
      " [103  39  12]\n",
      " [103  39  12]\n",
      " [101  40  12]\n",
      " [102  41  13]\n",
      " [102  41  13]\n",
      " [102  42  14]\n",
      " [102  42  14]\n",
      " [101  44  15]\n",
      " [102  45  16]\n",
      " [100  45  15]\n",
      " [100  45  15]\n",
      " [ 99  46  15]\n",
      " [100  47  16]\n",
      " [101  48  17]\n",
      " [100  49  18]\n",
      " [101  50  19]]\n"
     ]
    }
   ],
   "source": [
    "im = []\n",
    "for i in range(4):\n",
    "    line = np.hstack(imPredSmall[i * 4 : i * 4 + 4])\n",
    "    im.append(line)\n",
    "im = np.vstack(im)\n",
    "\n",
    "plt.imshow(im)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[57.897762  25.009315  13.612168 ]\n",
      " [66.08729   26.907211   8.88067  ]\n",
      " [69.00389   27.758347  11.015296 ]\n",
      " [70.15055   28.770735  10.605076 ]\n",
      " [70.53425   29.291563  10.464784 ]\n",
      " [70.68917   29.456049  10.257028 ]\n",
      " [71.94177   30.231426  10.917687 ]\n",
      " [72.27056   30.419395  10.4550085]\n",
      " [71.51132   29.563093   9.082475 ]\n",
      " [72.15798   29.843353   9.025834 ]\n",
      " [73.72932   31.031673   9.546621 ]\n",
      " [75.428246  32.41693   10.4217205]\n",
      " [77.03373   33.8509    11.353556 ]\n",
      " [77.970406  34.416065  11.671332 ]\n",
      " [79.39225   34.206852  12.282556 ]\n",
      " [82.69697   34.62863   12.094486 ]\n",
      " [84.53572   33.010067  11.073839 ]\n",
      " [84.30469   30.784533  10.619334 ]\n",
      " [84.00276   30.791286  10.564947 ]\n",
      " [83.7775    30.926058  10.551137 ]\n",
      " [84.06975   31.615284  11.109563 ]\n",
      " [84.41944   32.60601   11.861271 ]\n",
      " [84.59565   33.896187  12.461048 ]\n",
      " [84.34593   34.988556  12.496444 ]\n",
      " [83.306046  35.170532  11.713701 ]\n",
      " [83.86399   36.627613  12.806653 ]\n",
      " [82.845314  36.552315  12.533135 ]\n",
      " [82.28659   37.129692  12.605812 ]\n",
      " [82.22781   37.799706  13.267915 ]\n",
      " [82.42144   38.76429   15.220338 ]\n",
      " [74.96476   36.92839   16.95446  ]\n",
      " [65.21976   37.46312   20.149223 ]]\n"
     ]
    }
   ],
   "source": [
    "print(imPredSmall[0][0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 89  35  11]\n",
      " [ 89  35  11]\n",
      " [ 89  35  11]\n",
      " [ 90  36  12]\n",
      " [ 90  36  12]\n",
      " [ 90  36  10]\n",
      " [ 91  37  11]\n",
      " [ 91  37  11]\n",
      " [ 91  37   9]\n",
      " [ 91  37   9]\n",
      " [ 92  38  10]\n",
      " [ 93  39  11]\n",
      " [ 95  42  11]\n",
      " [ 96  43  12]\n",
      " [ 97  44  13]\n",
      " [ 97  44  13]\n",
      " [103  39  12]\n",
      " [103  39  12]\n",
      " [101  40  12]\n",
      " [102  41  13]\n",
      " [102  41  13]\n",
      " [102  42  14]\n",
      " [102  42  14]\n",
      " [101  44  15]\n",
      " [102  45  16]\n",
      " [100  45  15]\n",
      " [100  45  15]\n",
      " [ 99  46  15]\n",
      " [100  47  16]\n",
      " [101  48  17]\n",
      " [100  49  18]\n",
      " [101  50  19]]\n"
     ]
    }
   ],
   "source": [
    "print(imytestsmall[0][0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-33, -14,   6],\n",
       "       [-22, -12,  -2],\n",
       "       [-17, -11,   1],\n",
       "       [-17, -11,  -1],\n",
       "       [-17, -12,  -1],\n",
       "       [-17, -13,   1],\n",
       "       [-16, -12,   1],\n",
       "       [-15, -12,   1],\n",
       "       [-16, -14,   0],\n",
       "       [-17, -13,   0],\n",
       "       [-16, -13,   0],\n",
       "       [-15, -12,   0],\n",
       "       [-16, -13,   1],\n",
       "       [-15, -12,   1],\n",
       "       [-14, -11,   2],\n",
       "       [-17, -13,   0],\n",
       "       [-14, -13,   0],\n",
       "       [-14, -13,  -1],\n",
       "       [-14, -14,  -1],\n",
       "       [-13, -13,   0],\n",
       "       [-13, -13,   1],\n",
       "       [-13, -13,  -1],\n",
       "       [-12, -13,   1],\n",
       "       [-12, -13,   1],\n",
       "       [-12, -13,   2],\n",
       "       [-13, -13,   2],\n",
       "       [-13, -12,   1],\n",
       "       [-13, -12,   3],\n",
       "       [-13, -12,   2],\n",
       "       [-13, -12,   3],\n",
       "       [-25, -14,   3],\n",
       "       [-39, -14,   6]])"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "imPredSmall.astype('int')[1][0] - imytestsmall[1][0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10.778849283854166"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(np.abs(imPredSmall.astype('int') - imytestsmall[0:16]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6300.0"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "imXtestsmall.shape[0] / 16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "81.28673981843171"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(np.abs(Xtest- ytest))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "18.3308272427352"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(np.abs(Xtest.astype('int32') - ytest.astype('int32')))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[[ 79,  34,  11],\n",
       "         [ 79,  34,  11],\n",
       "         [ 79,  34,  11],\n",
       "         ...,\n",
       "         [ 31,  22,   7],\n",
       "         [ 31,  22,   7],\n",
       "         [ 31,  22,   7]],\n",
       "\n",
       "        [[ 79,  34,  11],\n",
       "         [ 79,  34,  11],\n",
       "         [ 80,  35,  12],\n",
       "         ...,\n",
       "         [ 30,  23,   7],\n",
       "         [ 31,  22,   7],\n",
       "         [ 31,  22,   7]],\n",
       "\n",
       "        [[ 80,  35,  12],\n",
       "         [ 80,  35,  12],\n",
       "         [ 81,  37,  12],\n",
       "         ...,\n",
       "         [ 31,  24,   8],\n",
       "         [ 30,  23,   7],\n",
       "         [ 30,  23,   7]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         ...,\n",
       "         [  5,   7,   6],\n",
       "         [  3,   7,   6],\n",
       "         [  2,   8,   6]],\n",
       "\n",
       "        [[  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         ...,\n",
       "         [ 11,  13,  12],\n",
       "         [ 14,  18,  17],\n",
       "         [ 14,  20,  18]],\n",
       "\n",
       "        [[  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         ...,\n",
       "         [ 16,  18,  15],\n",
       "         [ 23,  27,  26],\n",
       "         [ 25,  31,  29]]],\n",
       "\n",
       "\n",
       "       [[[ 29,  39,  28],\n",
       "         [ 38,  50,  40],\n",
       "         [  0,   6,   0],\n",
       "         ...,\n",
       "         [  3,   0,   0],\n",
       "         [  1,   0,   0],\n",
       "         [  0,   0,   0]],\n",
       "\n",
       "        [[ 52,  62,  51],\n",
       "         [ 55,  65,  56],\n",
       "         [  1,  13,   3],\n",
       "         ...,\n",
       "         [  1,   0,   0],\n",
       "         [  1,   0,   0],\n",
       "         [  0,   0,   0]],\n",
       "\n",
       "        [[ 73,  83,  74],\n",
       "         [ 68,  78,  69],\n",
       "         [  7,  17,   9],\n",
       "         ...,\n",
       "         [  1,   0,   0],\n",
       "         [  0,   0,   0],\n",
       "         [  0,   0,   2]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[  5,  46,  52],\n",
       "         [  0,  41,  45],\n",
       "         [  0,  42,  40],\n",
       "         ...,\n",
       "         [ 30,  51,  46],\n",
       "         [ 38,  58,  59],\n",
       "         [ 43,  62,  66]],\n",
       "\n",
       "        [[  1,  44,  51],\n",
       "         [  0,  44,  47],\n",
       "         [  1,  47,  45],\n",
       "         ...,\n",
       "         [ 30,  49,  43],\n",
       "         [ 37,  55,  55],\n",
       "         [ 35,  53,  57]],\n",
       "\n",
       "        [[  0,  42,  48],\n",
       "         [  1,  46,  51],\n",
       "         [  4,  53,  50],\n",
       "         ...,\n",
       "         [ 26,  45,  39],\n",
       "         [ 28,  46,  46],\n",
       "         [ 14,  32,  36]]],\n",
       "\n",
       "\n",
       "       [[[ 21,  42,  23],\n",
       "         [ 19,  40,  21],\n",
       "         [ 19,  40,  21],\n",
       "         ...,\n",
       "         [ 35,  53,  29],\n",
       "         [ 33,  51,  27],\n",
       "         [ 31,  49,  25]],\n",
       "\n",
       "        [[ 23,  44,  25],\n",
       "         [ 21,  42,  23],\n",
       "         [ 21,  42,  23],\n",
       "         ...,\n",
       "         [ 35,  53,  29],\n",
       "         [ 34,  52,  28],\n",
       "         [ 33,  51,  27]],\n",
       "\n",
       "        [[ 23,  44,  25],\n",
       "         [ 22,  43,  24],\n",
       "         [ 22,  43,  24],\n",
       "         ...,\n",
       "         [ 36,  54,  30],\n",
       "         [ 36,  54,  30],\n",
       "         [ 36,  54,  30]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[ 27,  45,  21],\n",
       "         [ 29,  47,  23],\n",
       "         [ 36,  39,  20],\n",
       "         ...,\n",
       "         [ 40,  54,  37],\n",
       "         [ 40,  54,  37],\n",
       "         [ 32,  46,  29]],\n",
       "\n",
       "        [[ 17,  42,  20],\n",
       "         [ 25,  53,  30],\n",
       "         [ 23,  38,  19],\n",
       "         ...,\n",
       "         [ 38,  56,  34],\n",
       "         [ 36,  53,  34],\n",
       "         [ 29,  46,  27]],\n",
       "\n",
       "        [[ 15,  49,  25],\n",
       "         [ 18,  50,  27],\n",
       "         [ 32,  49,  31],\n",
       "         ...,\n",
       "         [ 35,  55,  30],\n",
       "         [ 33,  52,  30],\n",
       "         [ 26,  45,  23]]],\n",
       "\n",
       "\n",
       "       ...,\n",
       "\n",
       "\n",
       "       [[[ 92, 225, 242],\n",
       "         [115, 240, 255],\n",
       "         [133, 253, 255],\n",
       "         ...,\n",
       "         [ 12,  16,  27],\n",
       "         [ 10,  17,  27],\n",
       "         [ 10,  17,  27]],\n",
       "\n",
       "        [[156, 254, 255],\n",
       "         [157, 252, 255],\n",
       "         [155, 250, 255],\n",
       "         ...,\n",
       "         [ 11,  15,  26],\n",
       "         [ 10,  17,  27],\n",
       "         [ 10,  17,  27]],\n",
       "\n",
       "        [[198, 253, 248],\n",
       "         [187, 246, 244],\n",
       "         [174, 240, 238],\n",
       "         ...,\n",
       "         [  9,  13,  24],\n",
       "         [  9,  16,  26],\n",
       "         [  9,  16,  26]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         ...,\n",
       "         [  3,  11,  14],\n",
       "         [  1,   9,  12],\n",
       "         [  0,   5,   8]],\n",
       "\n",
       "        [[  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         ...,\n",
       "         [  5,  12,  20],\n",
       "         [  6,  10,  19],\n",
       "         [  1,   5,  14]],\n",
       "\n",
       "        [[  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         ...,\n",
       "         [  3,   7,  18],\n",
       "         [  5,   9,  20],\n",
       "         [  5,   7,  20]]],\n",
       "\n",
       "\n",
       "       [[[ 87,  73,  62],\n",
       "         [ 86,  69,  59],\n",
       "         [ 83,  66,  56],\n",
       "         ...,\n",
       "         [ 68,  61,  53],\n",
       "         [ 68,  61,  53],\n",
       "         [ 68,  61,  53]],\n",
       "\n",
       "        [[ 78,  64,  51],\n",
       "         [ 78,  64,  51],\n",
       "         [ 79,  65,  54],\n",
       "         ...,\n",
       "         [ 67,  60,  52],\n",
       "         [ 67,  60,  52],\n",
       "         [ 66,  59,  51]],\n",
       "\n",
       "        [[103,  83,  72],\n",
       "         [ 96,  78,  66],\n",
       "         [ 88,  70,  58],\n",
       "         ...,\n",
       "         [ 68,  61,  53],\n",
       "         [ 67,  60,  52],\n",
       "         [ 67,  60,  52]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[166, 123, 106],\n",
       "         [145, 106,  89],\n",
       "         [116,  84,  69],\n",
       "         ...,\n",
       "         [169, 126,  92],\n",
       "         [169, 126,  92],\n",
       "         [169, 126,  92]],\n",
       "\n",
       "        [[172, 124, 104],\n",
       "         [158, 113,  94],\n",
       "         [149, 112,  94],\n",
       "         ...,\n",
       "         [170, 127,  93],\n",
       "         [169, 126,  92],\n",
       "         [169, 126,  92]],\n",
       "\n",
       "        [[167, 115,  94],\n",
       "         [152, 105,  85],\n",
       "         [171, 130, 112],\n",
       "         ...,\n",
       "         [170, 127,  93],\n",
       "         [169, 126,  92],\n",
       "         [169, 126,  92]]],\n",
       "\n",
       "\n",
       "       [[[211, 140,  36],\n",
       "         [215, 146,  45],\n",
       "         [206, 137,  44],\n",
       "         ...,\n",
       "         [240, 231, 250],\n",
       "         [244, 246, 255],\n",
       "         [223, 236, 244]],\n",
       "\n",
       "        [[236, 179,  90],\n",
       "         [239, 178,  95],\n",
       "         [236, 175,  95],\n",
       "         ...,\n",
       "         [228, 229, 233],\n",
       "         [216, 224, 226],\n",
       "         [189, 204, 201]],\n",
       "\n",
       "        [[217, 167,  94],\n",
       "         [217, 163,  93],\n",
       "         [223, 163, 100],\n",
       "         ...,\n",
       "         [ 54,  53,  35],\n",
       "         [ 41,  45,  20],\n",
       "         [ 34,  45,  13]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[232, 232, 230],\n",
       "         [232, 232, 230],\n",
       "         [232, 232, 230],\n",
       "         ...,\n",
       "         [226, 228, 227],\n",
       "         [220, 224, 223],\n",
       "         [227, 231, 232]],\n",
       "\n",
       "        [[234, 233, 238],\n",
       "         [234, 233, 238],\n",
       "         [234, 234, 236],\n",
       "         ...,\n",
       "         [230, 232, 231],\n",
       "         [228, 232, 231],\n",
       "         [230, 234, 235]],\n",
       "\n",
       "        [[232, 231, 239],\n",
       "         [232, 231, 237],\n",
       "         [233, 232, 238],\n",
       "         ...,\n",
       "         [229, 233, 232],\n",
       "         [229, 233, 234],\n",
       "         [229, 233, 234]]]])"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Xtest.astype('int32') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[[ 79,  34,  11],\n",
       "         [ 79,  34,  11],\n",
       "         [ 79,  34,  11],\n",
       "         ...,\n",
       "         [ 31,  22,   7],\n",
       "         [ 31,  22,   7],\n",
       "         [ 31,  22,   7]],\n",
       "\n",
       "        [[ 79,  34,  11],\n",
       "         [ 79,  34,  11],\n",
       "         [ 80,  35,  12],\n",
       "         ...,\n",
       "         [ 30,  23,   7],\n",
       "         [ 31,  22,   7],\n",
       "         [ 31,  22,   7]],\n",
       "\n",
       "        [[ 80,  35,  12],\n",
       "         [ 80,  35,  12],\n",
       "         [ 81,  37,  12],\n",
       "         ...,\n",
       "         [ 31,  24,   8],\n",
       "         [ 30,  23,   7],\n",
       "         [ 30,  23,   7]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         ...,\n",
       "         [  5,   7,   6],\n",
       "         [  3,   7,   6],\n",
       "         [  2,   8,   6]],\n",
       "\n",
       "        [[  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         ...,\n",
       "         [ 11,  13,  12],\n",
       "         [ 14,  18,  17],\n",
       "         [ 14,  20,  18]],\n",
       "\n",
       "        [[  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         [  1,   3,   0],\n",
       "         ...,\n",
       "         [ 16,  18,  15],\n",
       "         [ 23,  27,  26],\n",
       "         [ 25,  31,  29]]],\n",
       "\n",
       "\n",
       "       [[[ 29,  39,  28],\n",
       "         [ 38,  50,  40],\n",
       "         [  0,   6,   0],\n",
       "         ...,\n",
       "         [  3,   0,   0],\n",
       "         [  1,   0,   0],\n",
       "         [  0,   0,   0]],\n",
       "\n",
       "        [[ 52,  62,  51],\n",
       "         [ 55,  65,  56],\n",
       "         [  1,  13,   3],\n",
       "         ...,\n",
       "         [  1,   0,   0],\n",
       "         [  1,   0,   0],\n",
       "         [  0,   0,   0]],\n",
       "\n",
       "        [[ 73,  83,  74],\n",
       "         [ 68,  78,  69],\n",
       "         [  7,  17,   9],\n",
       "         ...,\n",
       "         [  1,   0,   0],\n",
       "         [  0,   0,   0],\n",
       "         [  0,   0,   2]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[  5,  46,  52],\n",
       "         [  0,  41,  45],\n",
       "         [  0,  42,  40],\n",
       "         ...,\n",
       "         [ 30,  51,  46],\n",
       "         [ 38,  58,  59],\n",
       "         [ 43,  62,  66]],\n",
       "\n",
       "        [[  1,  44,  51],\n",
       "         [  0,  44,  47],\n",
       "         [  1,  47,  45],\n",
       "         ...,\n",
       "         [ 30,  49,  43],\n",
       "         [ 37,  55,  55],\n",
       "         [ 35,  53,  57]],\n",
       "\n",
       "        [[  0,  42,  48],\n",
       "         [  1,  46,  51],\n",
       "         [  4,  53,  50],\n",
       "         ...,\n",
       "         [ 26,  45,  39],\n",
       "         [ 28,  46,  46],\n",
       "         [ 14,  32,  36]]],\n",
       "\n",
       "\n",
       "       [[[ 21,  42,  23],\n",
       "         [ 19,  40,  21],\n",
       "         [ 19,  40,  21],\n",
       "         ...,\n",
       "         [ 35,  53,  29],\n",
       "         [ 33,  51,  27],\n",
       "         [ 31,  49,  25]],\n",
       "\n",
       "        [[ 23,  44,  25],\n",
       "         [ 21,  42,  23],\n",
       "         [ 21,  42,  23],\n",
       "         ...,\n",
       "         [ 35,  53,  29],\n",
       "         [ 34,  52,  28],\n",
       "         [ 33,  51,  27]],\n",
       "\n",
       "        [[ 23,  44,  25],\n",
       "         [ 22,  43,  24],\n",
       "         [ 22,  43,  24],\n",
       "         ...,\n",
       "         [ 36,  54,  30],\n",
       "         [ 36,  54,  30],\n",
       "         [ 36,  54,  30]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[ 27,  45,  21],\n",
       "         [ 29,  47,  23],\n",
       "         [ 36,  39,  20],\n",
       "         ...,\n",
       "         [ 40,  54,  37],\n",
       "         [ 40,  54,  37],\n",
       "         [ 32,  46,  29]],\n",
       "\n",
       "        [[ 17,  42,  20],\n",
       "         [ 25,  53,  30],\n",
       "         [ 23,  38,  19],\n",
       "         ...,\n",
       "         [ 38,  56,  34],\n",
       "         [ 36,  53,  34],\n",
       "         [ 29,  46,  27]],\n",
       "\n",
       "        [[ 15,  49,  25],\n",
       "         [ 18,  50,  27],\n",
       "         [ 32,  49,  31],\n",
       "         ...,\n",
       "         [ 35,  55,  30],\n",
       "         [ 33,  52,  30],\n",
       "         [ 26,  45,  23]]],\n",
       "\n",
       "\n",
       "       ...,\n",
       "\n",
       "\n",
       "       [[[ 92, 225, 242],\n",
       "         [115, 240, 255],\n",
       "         [133, 253, 255],\n",
       "         ...,\n",
       "         [ 12,  16,  27],\n",
       "         [ 10,  17,  27],\n",
       "         [ 10,  17,  27]],\n",
       "\n",
       "        [[156, 254, 255],\n",
       "         [157, 252, 255],\n",
       "         [155, 250, 255],\n",
       "         ...,\n",
       "         [ 11,  15,  26],\n",
       "         [ 10,  17,  27],\n",
       "         [ 10,  17,  27]],\n",
       "\n",
       "        [[198, 253, 248],\n",
       "         [187, 246, 244],\n",
       "         [174, 240, 238],\n",
       "         ...,\n",
       "         [  9,  13,  24],\n",
       "         [  9,  16,  26],\n",
       "         [  9,  16,  26]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         ...,\n",
       "         [  3,  11,  14],\n",
       "         [  1,   9,  12],\n",
       "         [  0,   5,   8]],\n",
       "\n",
       "        [[  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         ...,\n",
       "         [  5,  12,  20],\n",
       "         [  6,  10,  19],\n",
       "         [  1,   5,  14]],\n",
       "\n",
       "        [[  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         [  4,   4,  12],\n",
       "         ...,\n",
       "         [  3,   7,  18],\n",
       "         [  5,   9,  20],\n",
       "         [  5,   7,  20]]],\n",
       "\n",
       "\n",
       "       [[[ 87,  73,  62],\n",
       "         [ 86,  69,  59],\n",
       "         [ 83,  66,  56],\n",
       "         ...,\n",
       "         [ 68,  61,  53],\n",
       "         [ 68,  61,  53],\n",
       "         [ 68,  61,  53]],\n",
       "\n",
       "        [[ 78,  64,  51],\n",
       "         [ 78,  64,  51],\n",
       "         [ 79,  65,  54],\n",
       "         ...,\n",
       "         [ 67,  60,  52],\n",
       "         [ 67,  60,  52],\n",
       "         [ 66,  59,  51]],\n",
       "\n",
       "        [[103,  83,  72],\n",
       "         [ 96,  78,  66],\n",
       "         [ 88,  70,  58],\n",
       "         ...,\n",
       "         [ 68,  61,  53],\n",
       "         [ 67,  60,  52],\n",
       "         [ 67,  60,  52]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[166, 123, 106],\n",
       "         [145, 106,  89],\n",
       "         [116,  84,  69],\n",
       "         ...,\n",
       "         [169, 126,  92],\n",
       "         [169, 126,  92],\n",
       "         [169, 126,  92]],\n",
       "\n",
       "        [[172, 124, 104],\n",
       "         [158, 113,  94],\n",
       "         [149, 112,  94],\n",
       "         ...,\n",
       "         [170, 127,  93],\n",
       "         [169, 126,  92],\n",
       "         [169, 126,  92]],\n",
       "\n",
       "        [[167, 115,  94],\n",
       "         [152, 105,  85],\n",
       "         [171, 130, 112],\n",
       "         ...,\n",
       "         [170, 127,  93],\n",
       "         [169, 126,  92],\n",
       "         [169, 126,  92]]],\n",
       "\n",
       "\n",
       "       [[[211, 140,  36],\n",
       "         [215, 146,  45],\n",
       "         [206, 137,  44],\n",
       "         ...,\n",
       "         [240, 231, 250],\n",
       "         [244, 246, 255],\n",
       "         [223, 236, 244]],\n",
       "\n",
       "        [[236, 179,  90],\n",
       "         [239, 178,  95],\n",
       "         [236, 175,  95],\n",
       "         ...,\n",
       "         [228, 229, 233],\n",
       "         [216, 224, 226],\n",
       "         [189, 204, 201]],\n",
       "\n",
       "        [[217, 167,  94],\n",
       "         [217, 163,  93],\n",
       "         [223, 163, 100],\n",
       "         ...,\n",
       "         [ 54,  53,  35],\n",
       "         [ 41,  45,  20],\n",
       "         [ 34,  45,  13]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[232, 232, 230],\n",
       "         [232, 232, 230],\n",
       "         [232, 232, 230],\n",
       "         ...,\n",
       "         [226, 228, 227],\n",
       "         [220, 224, 223],\n",
       "         [227, 231, 232]],\n",
       "\n",
       "        [[234, 233, 238],\n",
       "         [234, 233, 238],\n",
       "         [234, 234, 236],\n",
       "         ...,\n",
       "         [230, 232, 231],\n",
       "         [228, 232, 231],\n",
       "         [230, 234, 235]],\n",
       "\n",
       "        [[232, 231, 239],\n",
       "         [232, 231, 237],\n",
       "         [233, 232, 238],\n",
       "         ...,\n",
       "         [229, 233, 232],\n",
       "         [229, 233, 234],\n",
       "         [229, 233, 234]]]], dtype=uint8)"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Xtest"
   ]
  },
  {
   "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}