3DRecognition.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "source": [
    "### Imports"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 666,
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "import trimesh\n",
    "import random\n",
    "import math\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import tensorflow as tf\n",
    "import keras\n",
    "import keras.models\n",
    "\n",
    "from PIL import Image\n",
    "from random import seed\n",
    "from random import random\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "from sklearn.metrics import ConfusionMatrixDisplay\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "import keyboard\n",
    "import seaborn as sns"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Read data"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "outputs": [],
   "source": [
    "start = 1"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 264,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   },
   "outputs": [],
   "source": [
    "def read_off(file):\n",
    "    return trimesh.load(file).sample(8192)\n",
    "\n",
    "\n",
    "def plot3D(data, save):\n",
    "    global start\n",
    "    fig = plt.figure(figsize=(5, 5))\n",
    "    ax = fig.add_subplot(111, projection=\"3d\")\n",
    "    ax.scatter(data[:, 0], data[:, 1], data[:, 2])\n",
    "    ax.view_init(30, 200)\n",
    "    if save:\n",
    "        start += 1\n",
    "        plt.savefig('images/' + str(start) + '.png')\n",
    "\n",
    "def plot2DBW(data, save):\n",
    "    global start\n",
    "    plt.imshow(data[1],'gray')\n",
    "    if save:\n",
    "        start += 1\n",
    "        plt.savefig('images/' + str(start) + '.png')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 chair_0292.off\n",
      "1 chair_0603.off\n",
      "2 chair_0098.off\n",
      "3 chair_0803.off\n",
      "4 chair_0331.off\n",
      "5 chair_0580.off\n",
      "6 chair_0420.off\n",
      "7 chair_0409.off\n",
      "8 chair_0315.off\n",
      "9 chair_0285.off\n",
      "10 chair_0018.off\n",
      "11 chair_0125.off\n",
      "12 chair_0451.off\n",
      "13 chair_0425.off\n",
      "14 chair_0242.off\n",
      "15 chair_0606.off\n",
      "16 chair_0255.off\n",
      "17 chair_0769.off\n",
      "18 chair_0849.off\n",
      "19 chair_0640.off\n",
      "20 chair_0762.off\n",
      "21 chair_0032.off\n",
      "22 chair_0517.off\n",
      "23 chair_0245.off\n",
      "24 chair_0369.off\n",
      "25 chair_0504.off\n",
      "26 chair_0110.off\n",
      "27 chair_0262.off\n",
      "28 chair_0365.off\n",
      "29 chair_0567.off\n",
      "30 chair_0714.off\n",
      "31 chair_0116.off\n",
      "32 chair_0850.off\n",
      "33 chair_0014.off\n",
      "34 chair_0367.off\n",
      "35 chair_0186.off\n",
      "36 chair_0623.off\n",
      "37 chair_0586.off\n",
      "38 chair_0824.off\n",
      "39 chair_0460.off\n",
      "40 chair_0299.off\n",
      "41 chair_0598.off\n",
      "42 chair_0697.off\n",
      "43 chair_0826.off\n",
      "44 chair_0020.off\n",
      "45 chair_0660.off\n",
      "46 chair_0056.off\n",
      "47 chair_0424.off\n",
      "48 chair_0495.off\n",
      "49 chair_0398.off\n",
      "50 chair_0791.off\n",
      "51 chair_0179.off\n",
      "52 chair_0079.off\n",
      "53 chair_0140.off\n",
      "54 chair_0224.off\n",
      "55 chair_0337.off\n",
      "56 chair_0452.off\n",
      "57 chair_0348.off\n",
      "58 chair_0488.off\n",
      "59 chair_0733.off\n",
      "60 chair_0827.off\n",
      "61 chair_0150.off\n",
      "62 chair_0652.off\n",
      "63 chair_0663.off\n",
      "64 chair_0099.off\n",
      "65 chair_0723.off\n",
      "66 chair_0073.off\n",
      "67 chair_0710.off\n",
      "68 chair_0075.off\n",
      "69 chair_0715.off\n",
      "70 chair_0156.off\n",
      "71 chair_0025.off\n",
      "72 chair_0731.off\n",
      "73 chair_0394.off\n",
      "74 chair_0785.off\n",
      "75 chair_0799.off\n",
      "76 chair_0823.off\n",
      "77 chair_0003.off\n",
      "78 chair_0072.off\n",
      "79 chair_0147.off\n",
      "80 chair_0089.off\n",
      "81 chair_0048.off\n",
      "82 chair_0024.off\n",
      "83 chair_0583.off\n",
      "84 chair_0675.off\n",
      "85 chair_0358.off\n",
      "86 chair_0151.off\n",
      "87 chair_0228.off\n",
      "88 chair_0203.off\n",
      "89 chair_0869.off\n",
      "90 chair_0066.off\n",
      "91 chair_0013.off\n",
      "92 chair_0158.off\n",
      "93 chair_0742.off\n",
      "94 chair_0796.off\n",
      "95 chair_0848.off\n",
      "96 chair_0240.off\n",
      "97 chair_0354.off\n",
      "98 chair_0458.off\n",
      "99 chair_0293.off\n",
      "100 chair_0722.off\n",
      "101 chair_0886.off\n",
      "102 chair_0145.off\n",
      "103 chair_0533.off\n",
      "104 chair_0162.off\n",
      "105 chair_0596.off\n",
      "106 chair_0825.off\n",
      "107 chair_0253.off\n",
      "108 chair_0763.off\n",
      "109 chair_0741.off\n",
      "110 chair_0113.off\n",
      "111 chair_0518.off\n",
      "112 chair_0055.off\n",
      "113 chair_0859.off\n",
      "114 chair_0882.off\n",
      "115 chair_0615.off\n",
      "116 chair_0511.off\n",
      "117 chair_0212.off\n",
      "118 chair_0256.off\n",
      "119 chair_0172.off\n",
      "120 chair_0264.off\n",
      "121 chair_0669.off\n",
      "122 chair_0593.off\n",
      "123 chair_0672.off\n",
      "124 chair_0578.off\n",
      "125 chair_0473.off\n",
      "126 chair_0459.off\n",
      "127 chair_0772.off\n",
      "128 chair_0132.off\n",
      "129 chair_0193.off\n",
      "130 chair_0249.off\n",
      "131 chair_0734.off\n",
      "132 chair_0555.off\n",
      "133 chair_0065.off\n",
      "134 chair_0484.off\n",
      "135 chair_0857.off\n",
      "136 chair_0647.off\n",
      "137 chair_0626.off\n",
      "138 chair_0050.off\n",
      "139 chair_0231.off\n",
      "140 chair_0474.off\n",
      "141 chair_0455.off\n",
      "142 chair_0301.off\n",
      "143 chair_0534.off\n",
      "144 chair_0342.off\n",
      "145 chair_0263.off\n",
      "146 chair_0686.off\n",
      "147 chair_0873.off\n",
      "148 chair_0443.off\n",
      "149 chair_0434.off\n",
      "150 chair_0487.off\n",
      "151 chair_0696.off\n",
      "152 chair_0416.off\n",
      "153 chair_0215.off\n",
      "154 chair_0092.off\n",
      "155 chair_0010.off\n",
      "156 chair_0311.off\n",
      "157 chair_0017.off\n",
      "158 chair_0370.off\n",
      "159 chair_0679.off\n",
      "160 chair_0592.off\n",
      "161 chair_0876.off\n",
      "162 chair_0239.off\n",
      "163 chair_0419.off\n",
      "164 chair_0564.off\n",
      "165 chair_0538.off\n",
      "166 chair_0317.off\n",
      "167 chair_0889.off\n",
      "168 chair_0834.off\n",
      "169 chair_0226.off\n",
      "170 chair_0407.off\n",
      "171 chair_0807.off\n",
      "172 chair_0683.off\n",
      "173 chair_0246.off\n",
      "174 chair_0019.off\n",
      "175 chair_0786.off\n",
      "176 chair_0818.off\n",
      "177 chair_0554.off\n",
      "178 chair_0221.off\n",
      "179 chair_0752.off\n",
      "180 chair_0489.off\n",
      "181 chair_0657.off\n",
      "182 chair_0054.off\n",
      "183 chair_0314.off\n",
      "184 chair_0462.off\n",
      "185 chair_0343.off\n",
      "186 chair_0036.off\n",
      "187 chair_0472.off\n",
      "188 chair_0321.off\n",
      "189 chair_0840.off\n",
      "190 chair_0155.off\n",
      "191 chair_0608.off\n",
      "192 chair_0142.off\n",
      "193 chair_0015.off\n",
      "194 chair_0736.off\n",
      "195 chair_0189.off\n",
      "196 chair_0498.off\n",
      "197 chair_0562.off\n",
      "198 chair_0497.off\n",
      "199 chair_0103.off\n",
      "200 chair_0180.off\n",
      "201 chair_0878.off\n",
      "202 chair_0261.off\n",
      "203 chair_0632.off\n",
      "204 chair_0757.off\n",
      "205 chair_0559.off\n",
      "206 chair_0674.off\n",
      "207 chair_0356.off\n",
      "208 chair_0885.off\n",
      "209 chair_0695.off\n",
      "210 chair_0745.off\n",
      "211 chair_0523.off\n",
      "212 chair_0478.off\n",
      "213 chair_0159.off\n",
      "214 chair_0081.off\n",
      "215 chair_0871.off\n",
      "216 chair_0447.off\n",
      "217 chair_0133.off\n",
      "218 chair_0719.off\n",
      "219 chair_0808.off\n",
      "220 chair_0879.off\n",
      "221 chair_0453.off\n",
      "222 chair_0289.off\n",
      "223 chair_0645.off\n",
      "224 chair_0112.off\n",
      "225 chair_0449.off\n",
      "226 chair_0492.off\n",
      "227 chair_0627.off\n",
      "228 chair_0793.off\n",
      "229 chair_0527.off\n",
      "230 chair_0653.off\n",
      "231 chair_0045.off\n",
      "232 chair_0875.off\n",
      "233 chair_0549.off\n",
      "234 chair_0735.off\n",
      "235 chair_0512.off\n",
      "236 chair_0330.off\n",
      "237 chair_0713.off\n",
      "238 chair_0235.off\n",
      "239 chair_0599.off\n",
      "240 chair_0432.off\n",
      "241 chair_0102.off\n",
      "242 chair_0177.off\n",
      "243 chair_0022.off\n",
      "244 chair_0316.off\n",
      "245 chair_0551.off\n",
      "246 chair_0217.off\n",
      "247 chair_0664.off\n",
      "248 chair_0839.off\n",
      "249 chair_0649.off\n",
      "250 chair_0863.off\n",
      "251 chair_0171.off\n",
      "252 chair_0819.off\n",
      "253 chair_0728.off\n",
      "254 chair_0207.off\n",
      "255 chair_0196.off\n",
      "256 chair_0468.off\n",
      "257 chair_0491.off\n",
      "258 chair_0528.off\n",
      "259 chair_0332.off\n",
      "260 chair_0111.off\n",
      "261 chair_0327.off\n",
      "262 chair_0114.off\n",
      "263 chair_0846.off\n",
      "264 chair_0643.off\n",
      "265 chair_0400.off\n",
      "266 chair_0738.off\n",
      "267 chair_0693.off\n",
      "268 chair_0732.off\n",
      "269 chair_0510.off\n",
      "270 chair_0059.off\n",
      "271 chair_0199.off\n",
      "272 chair_0764.off\n",
      "273 chair_0770.off\n",
      "274 chair_0844.off\n",
      "275 chair_0187.off\n",
      "276 chair_0509.off\n",
      "277 chair_0665.off\n",
      "278 chair_0326.off\n",
      "279 chair_0579.off\n",
      "280 chair_0130.off\n",
      "281 chair_0101.off\n",
      "282 chair_0404.off\n",
      "283 chair_0590.off\n",
      "284 chair_0494.off\n",
      "285 chair_0749.off\n",
      "286 chair_0884.off\n",
      "287 chair_0279.off\n",
      "288 chair_0535.off\n",
      "289 chair_0861.off\n",
      "290 chair_0323.off\n",
      "291 chair_0082.off\n",
      "292 chair_0234.off\n",
      "293 chair_0211.off\n",
      "294 chair_0440.off\n",
      "295 chair_0501.off\n",
      "296 chair_0373.off\n",
      "297 chair_0532.off\n",
      "298 chair_0856.off\n",
      "299 chair_0411.off\n",
      "300 chair_0148.off\n",
      "301 chair_0832.off\n",
      "302 chair_0670.off\n",
      "303 chair_0706.off\n",
      "304 chair_0168.off\n",
      "305 chair_0122.off\n",
      "306 chair_0257.off\n",
      "307 chair_0353.off\n",
      "308 chair_0573.off\n",
      "309 chair_0537.off\n",
      "310 chair_0300.off\n",
      "311 chair_0278.off\n",
      "312 chair_0457.off\n",
      "313 chair_0119.off\n",
      "314 chair_0874.off\n",
      "315 chair_0071.off\n",
      "316 chair_0095.off\n",
      "317 chair_0726.off\n",
      "318 chair_0413.off\n",
      "319 chair_0439.off\n",
      "320 chair_0777.off\n",
      "321 chair_0333.off\n",
      "322 chair_0389.off\n",
      "323 chair_0206.off\n",
      "324 chair_0220.off\n",
      "325 chair_0146.off\n",
      "326 chair_0691.off\n",
      "327 chair_0391.off\n",
      "328 chair_0479.off\n",
      "329 chair_0625.off\n",
      "330 chair_0530.off\n",
      "331 chair_0012.off\n",
      "332 chair_0805.off\n",
      "333 chair_0496.off\n",
      "334 chair_0406.off\n",
      "335 chair_0393.off\n",
      "336 chair_0390.off\n",
      "337 chair_0792.off\n",
      "338 chair_0233.off\n",
      "339 chair_0378.off\n",
      "340 chair_0174.off\n",
      "341 chair_0144.off\n",
      "342 chair_0553.off\n",
      "343 chair_0570.off\n",
      "344 chair_0499.off\n",
      "345 chair_0341.off\n",
      "346 chair_0650.off\n",
      "347 chair_0288.off\n",
      "348 chair_0236.off\n",
      "349 chair_0388.off\n",
      "350 chair_0318.off\n",
      "351 chair_0308.off\n",
      "352 chair_0104.off\n",
      "353 chair_0415.off\n",
      "354 chair_0408.off\n",
      "355 chair_0476.off\n",
      "356 chair_0867.off\n",
      "357 chair_0587.off\n",
      "358 chair_0274.off\n",
      "359 chair_0800.off\n",
      "360 chair_0604.off\n",
      "361 chair_0868.off\n",
      "362 chair_0536.off\n",
      "363 chair_0558.off\n",
      "364 chair_0853.off\n",
      "365 chair_0837.off\n",
      "366 chair_0568.off\n",
      "367 chair_0395.off\n",
      "368 chair_0160.off\n",
      "369 chair_0646.off\n",
      "370 chair_0338.off\n",
      "371 chair_0817.off\n",
      "372 chair_0164.off\n",
      "373 chair_0754.off\n",
      "374 chair_0561.off\n",
      "375 chair_0387.off\n",
      "376 chair_0282.off\n",
      "377 chair_0438.off\n",
      "378 chair_0724.off\n",
      "379 chair_0638.off\n",
      "380 chair_0781.off\n",
      "381 chair_0397.off\n",
      "382 chair_0477.off\n",
      "383 chair_0600.off\n",
      "384 chair_0436.off\n",
      "385 chair_0324.off\n",
      "386 chair_0730.off\n",
      "387 chair_0613.off\n",
      "388 chair_0096.off\n",
      "389 chair_0708.off\n",
      "390 chair_0507.off\n",
      "391 chair_0352.off\n",
      "392 chair_0229.off\n",
      "393 chair_0232.off\n",
      "394 chair_0392.off\n",
      "395 chair_0654.off\n",
      "396 chair_0872.off\n",
      "397 chair_0456.off\n",
      "398 chair_0173.off\n",
      "399 chair_0572.off\n",
      "400 chair_0127.off\n",
      "401 chair_0712.off\n",
      "402 chair_0085.off\n",
      "403 chair_0141.off\n",
      "404 chair_0667.off\n",
      "405 chair_0655.off\n",
      "406 chair_0254.off\n",
      "407 chair_0423.off\n",
      "408 chair_0202.off\n",
      "409 chair_0505.off\n",
      "410 chair_0771.off\n",
      "411 chair_0820.off\n",
      "412 chair_0426.off\n",
      "413 chair_0284.off\n",
      "414 chair_0184.off\n",
      "415 chair_0610.off\n",
      "416 chair_0280.off\n",
      "417 chair_0776.off\n",
      "418 chair_0379.off\n",
      "419 chair_0485.off\n",
      "420 chair_0128.off\n",
      "421 chair_0192.off\n",
      "422 chair_0822.off\n",
      "423 chair_0514.off\n",
      "424 chair_0773.off\n",
      "425 chair_0810.off\n",
      "426 chair_0175.off\n",
      "427 chair_0176.off\n",
      "428 chair_0688.off\n",
      "429 chair_0149.off\n",
      "430 chair_0816.off\n",
      "431 chair_0108.off\n",
      "432 chair_0143.off\n",
      "433 chair_0790.off\n",
      "434 chair_0023.off\n",
      "435 chair_0692.off\n",
      "436 chair_0243.off\n",
      "437 chair_0322.off\n",
      "438 chair_0068.off\n",
      "439 chair_0298.off\n",
      "440 chair_0247.off\n",
      "441 chair_0611.off\n",
      "442 chair_0349.off\n",
      "443 chair_0690.off\n",
      "444 chair_0428.off\n",
      "445 chair_0556.off\n",
      "446 chair_0581.off\n",
      "447 chair_0183.off\n",
      "448 chair_0466.off\n",
      "449 chair_0563.off\n",
      "450 chair_0325.off\n",
      "451 chair_0662.off\n",
      "452 chair_0531.off\n",
      "453 chair_0618.off\n",
      "454 chair_0313.off\n",
      "455 chair_0361.off\n",
      "456 chair_0401.off\n",
      "457 chair_0524.off\n",
      "458 chair_0060.off\n",
      "459 chair_0541.off\n",
      "460 chair_0344.off\n",
      "461 chair_0469.off\n",
      "462 chair_0814.off\n",
      "463 chair_0765.off\n",
      "464 chair_0634.off\n",
      "465 chair_0297.off\n",
      "466 chair_0481.off\n",
      "467 chair_0039.off\n",
      "468 chair_0360.off\n",
      "469 chair_0450.off\n",
      "470 chair_0368.off\n",
      "471 chair_0709.off\n",
      "472 chair_0153.off\n",
      "473 chair_0624.off\n",
      "474 chair_0107.off\n",
      "475 chair_0165.off\n",
      "476 chair_0794.off\n",
      "477 chair_0673.off\n",
      "478 chair_0412.off\n",
      "479 chair_0870.off\n",
      "480 chair_0258.off\n",
      "481 chair_0571.off\n",
      "482 chair_0656.off\n",
      "483 chair_0560.off\n",
      "484 chair_0828.off\n",
      "485 chair_0214.off\n",
      "486 chair_0329.off\n",
      "487 chair_0758.off\n",
      "488 chair_0801.off\n",
      "489 chair_0716.off\n",
      "490 chair_0483.off\n",
      "491 chair_0266.off\n",
      "492 chair_0812.off\n",
      "493 chair_0454.off\n",
      "494 chair_0087.off\n",
      "495 chair_0845.off\n",
      "496 chair_0100.off\n",
      "497 chair_0093.off\n",
      "498 chair_0005.off\n",
      "499 chair_0700.off\n",
      "500 chair_0355.off\n",
      "501 chair_0727.off\n",
      "502 chair_0250.off\n",
      "503 chair_0500.off\n",
      "504 chair_0062.off\n",
      "505 chair_0682.off\n",
      "506 chair_0287.off\n",
      "507 chair_0881.off\n",
      "508 chair_0547.off\n",
      "509 chair_0070.off\n",
      "510 chair_0865.off\n",
      "511 chair_0422.off\n",
      "512 chair_0506.off\n",
      "513 chair_0740.off\n",
      "514 chair_0833.off\n",
      "515 chair_0067.off\n",
      "516 chair_0617.off\n",
      "517 chair_0222.off\n",
      "518 chair_0047.off\n",
      "519 chair_0557.off\n",
      "520 chair_0435.off\n",
      "521 chair_0357.off\n",
      "522 chair_0376.off\n",
      "523 chair_0304.off\n",
      "524 chair_0272.off\n",
      "525 chair_0078.off\n",
      "526 chair_0516.off\n",
      "527 chair_0205.off\n",
      "528 chair_0223.off\n",
      "529 chair_0034.off\n",
      "530 chair_0658.off\n",
      "531 chair_0380.off\n",
      "532 chair_0703.off\n",
      "533 chair_0021.off\n",
      "534 chair_0157.off\n",
      "535 chair_0521.off\n",
      "536 chair_0806.off\n",
      "537 chair_0403.off\n",
      "538 chair_0381.off\n",
      "539 chair_0350.off\n",
      "540 chair_0305.off\n",
      "541 chair_0238.off\n",
      "542 chair_0644.off\n",
      "543 chair_0842.off\n",
      "544 chair_0748.off\n",
      "545 chair_0746.off\n",
      "546 chair_0046.off\n",
      "547 chair_0129.off\n",
      "548 chair_0739.off\n",
      "549 chair_0835.off\n",
      "550 chair_0126.off\n",
      "551 chair_0641.off\n",
      "552 chair_0543.off\n",
      "553 chair_0009.off\n",
      "554 chair_0026.off\n",
      "555 chair_0427.off\n",
      "556 chair_0004.off\n",
      "557 chair_0339.off\n",
      "558 chair_0582.off\n",
      "559 chair_0707.off\n",
      "560 chair_0383.off\n",
      "561 chair_0811.off\n",
      "562 chair_0591.off\n",
      "563 chair_0760.off\n",
      "564 chair_0084.off\n",
      "565 chair_0522.off\n",
      "566 chair_0668.off\n",
      "567 chair_0091.off\n",
      "568 chair_0248.off\n",
      "569 chair_0386.off\n",
      "570 chair_0259.off\n",
      "571 chair_0569.off\n",
      "572 chair_0195.off\n",
      "573 chair_0011.off\n",
      "574 chair_0852.off\n",
      "575 chair_0502.off\n",
      "576 chair_0463.off\n",
      "577 chair_0589.off\n",
      "578 chair_0545.off\n",
      "579 chair_0503.off\n",
      "580 chair_0028.off\n",
      "581 chair_0666.off\n",
      "582 chair_0198.off\n",
      "583 chair_0374.off\n",
      "584 chair_0041.off\n",
      "585 chair_0795.off\n",
      "586 chair_0756.off\n",
      "587 chair_0753.off\n",
      "588 chair_0437.off\n",
      "589 chair_0838.off\n",
      "590 chair_0320.off\n",
      "591 chair_0431.off\n",
      "592 chair_0843.off\n",
      "593 chair_0862.off\n",
      "594 chair_0766.off\n",
      "595 chair_0594.off\n",
      "596 chair_0001.off\n",
      "597 chair_0334.off\n",
      "598 chair_0346.off\n",
      "599 chair_0135.off\n",
      "600 chair_0251.off\n",
      "601 chair_0362.off\n",
      "602 chair_0744.off\n",
      "603 chair_0124.off\n",
      "604 chair_0821.off\n",
      "605 chair_0270.off\n",
      "606 chair_0090.off\n",
      "607 chair_0513.off\n",
      "608 chair_0830.off\n",
      "609 chair_0271.off\n",
      "610 chair_0191.off\n",
      "611 chair_0414.off\n",
      "612 chair_0759.off\n",
      "613 chair_0520.off\n",
      "614 chair_0219.off\n",
      "615 chair_0671.off\n",
      "616 chair_0120.off\n",
      "617 chair_0678.off\n",
      "618 chair_0064.off\n",
      "619 chair_0860.off\n",
      "620 chair_0614.off\n",
      "621 chair_0008.off\n",
      "622 chair_0218.off\n",
      "623 chair_0134.off\n",
      "624 chair_0841.off\n",
      "625 chair_0384.off\n",
      "626 chair_0444.off\n",
      "627 chair_0789.off\n",
      "628 chair_0063.off\n",
      "629 chair_0661.off\n",
      "630 chair_0775.off\n",
      "631 chair_0778.off\n",
      "632 chair_0628.off\n",
      "633 chair_0847.off\n",
      "634 chair_0585.off\n",
      "635 chair_0363.off\n",
      "636 chair_0815.off\n",
      "637 chair_0074.off\n",
      "638 chair_0550.off\n",
      "639 chair_0442.off\n",
      "640 chair_0161.off\n",
      "641 chair_0866.off\n",
      "642 chair_0225.off\n",
      "643 chair_0694.off\n",
      "644 chair_0464.off\n",
      "645 chair_0053.off\n",
      "646 chair_0680.off\n",
      "647 chair_0037.off\n",
      "648 chair_0761.off\n",
      "649 chair_0782.off\n",
      "650 chair_0286.off\n",
      "651 chair_0131.off\n",
      "652 chair_0080.off\n",
      "653 chair_0597.off\n",
      "654 chair_0448.off\n",
      "655 chair_0260.off\n",
      "656 chair_0787.off\n",
      "657 chair_0417.off\n",
      "658 chair_0802.off\n",
      "659 chair_0399.off\n",
      "660 chair_0565.off\n",
      "661 chair_0629.off\n",
      "662 chair_0637.off\n",
      "663 chair_0490.off\n",
      "664 chair_0750.off\n",
      "665 chair_0402.off\n",
      "666 chair_0601.off\n",
      "667 chair_0123.off\n",
      "668 chair_0575.off\n",
      "669 chair_0701.off\n",
      "670 chair_0312.off\n",
      "671 chair_0115.off\n",
      "672 chair_0539.off\n",
      "673 chair_0804.off\n",
      "674 chair_0684.off\n",
      "675 chair_0366.off\n",
      "676 chair_0051.off\n",
      "677 chair_0405.off\n",
      "678 chair_0268.off\n",
      "679 chair_0208.off\n",
      "680 chair_0540.off\n",
      "681 chair_0642.off\n",
      "682 chair_0636.off\n",
      "683 chair_0767.off\n",
      "684 chair_0552.off\n",
      "685 chair_0335.off\n",
      "686 chair_0396.off\n",
      "687 chair_0797.off\n",
      "688 chair_0681.off\n",
      "689 chair_0086.off\n",
      "690 chair_0276.off\n",
      "691 chair_0676.off\n",
      "692 chair_0465.off\n",
      "693 chair_0273.off\n",
      "694 chair_0525.off\n",
      "695 chair_0328.off\n",
      "696 chair_0747.off\n",
      "697 chair_0621.off\n",
      "698 chair_0043.off\n",
      "699 chair_0283.off\n",
      "700 chair_0470.off\n",
      "701 chair_0574.off\n",
      "702 chair_0296.off\n",
      "703 chair_0044.off\n",
      "704 chair_0364.off\n",
      "705 chair_0336.off\n",
      "706 chair_0007.off\n",
      "707 chair_0659.off\n",
      "708 chair_0698.off\n",
      "709 chair_0577.off\n",
      "710 chair_0088.off\n",
      "711 chair_0858.off\n",
      "712 chair_0042.off\n",
      "713 chair_0566.off\n",
      "714 chair_0006.off\n",
      "715 chair_0648.off\n",
      "716 chair_0138.off\n",
      "717 chair_0319.off\n",
      "718 chair_0265.off\n",
      "719 chair_0548.off\n",
      "720 chair_0309.off\n",
      "721 chair_0371.off\n",
      "722 chair_0631.off\n",
      "723 chair_0241.off\n",
      "724 chair_0139.off\n",
      "725 chair_0216.off\n",
      "726 chair_0166.off\n",
      "727 chair_0181.off\n",
      "728 chair_0864.off\n",
      "729 chair_0788.off\n",
      "730 chair_0382.off\n",
      "731 chair_0711.off\n",
      "732 chair_0446.off\n",
      "733 chair_0883.off\n",
      "734 chair_0351.off\n",
      "735 chair_0471.off\n",
      "736 chair_0526.off\n",
      "737 chair_0831.off\n",
      "738 chair_0029.off\n",
      "739 chair_0275.off\n",
      "740 chair_0836.off\n",
      "741 chair_0410.off\n",
      "742 chair_0097.off\n",
      "743 chair_0295.off\n",
      "744 chair_0705.off\n",
      "745 chair_0213.off\n",
      "746 chair_0169.off\n",
      "747 chair_0152.off\n",
      "748 chair_0576.off\n",
      "749 chair_0888.off\n",
      "750 chair_0651.off\n",
      "751 chair_0493.off\n",
      "752 chair_0303.off\n",
      "753 chair_0813.off\n",
      "754 chair_0375.off\n",
      "755 chair_0290.off\n",
      "756 chair_0118.off\n",
      "757 chair_0377.off\n",
      "758 chair_0630.off\n",
      "759 chair_0720.off\n",
      "760 chair_0069.off\n",
      "761 chair_0685.off\n",
      "762 chair_0188.off\n",
      "763 chair_0197.off\n",
      "764 chair_0137.off\n",
      "765 chair_0016.off\n",
      "766 chair_0281.off\n",
      "767 chair_0302.off\n",
      "768 chair_0154.off\n",
      "769 chair_0372.off\n",
      "770 chair_0210.off\n",
      "771 chair_0717.off\n",
      "772 chair_0163.off\n",
      "773 chair_0441.off\n",
      "774 chair_0345.off\n",
      "775 chair_0609.off\n",
      "776 chair_0310.off\n",
      "777 chair_0607.off\n",
      "778 chair_0779.off\n",
      "779 chair_0704.off\n",
      "780 chair_0200.off\n",
      "781 chair_0546.off\n",
      "782 chair_0612.off\n",
      "783 chair_0774.off\n",
      "784 chair_0136.off\n",
      "785 chair_0252.off\n",
      "786 chair_0030.off\n",
      "787 chair_0291.off\n",
      "788 chair_0306.off\n",
      "789 chair_0461.off\n",
      "790 chair_0340.off\n",
      "791 chair_0616.off\n",
      "792 chair_0049.off\n",
      "793 chair_0809.off\n",
      "794 chair_0721.off\n",
      "795 chair_0486.off\n",
      "796 chair_0542.off\n",
      "797 chair_0307.off\n",
      "798 chair_0031.off\n",
      "799 chair_0294.off\n",
      "800 chair_0783.off\n",
      "801 chair_0052.off\n",
      "802 chair_0002.off\n",
      "803 chair_0106.off\n",
      "804 chair_0076.off\n",
      "805 chair_0751.off\n",
      "806 chair_0689.off\n",
      "807 chair_0277.off\n",
      "808 chair_0121.off\n",
      "809 chair_0702.off\n",
      "810 chair_0347.off\n",
      "811 chair_0433.off\n",
      "812 chair_0619.off\n",
      "813 chair_0105.off\n",
      "814 chair_0209.off\n",
      "815 chair_0482.off\n",
      "816 chair_0033.off\n",
      "817 chair_0729.off\n",
      "818 chair_0083.off\n",
      "819 chair_0755.off\n",
      "820 chair_0677.off\n",
      "821 chair_0854.off\n",
      "822 chair_0639.off\n",
      "823 chair_0602.off\n",
      "824 chair_0227.off\n",
      "825 chair_0887.off\n",
      "826 chair_0418.off\n",
      "827 chair_0190.off\n",
      "828 chair_0737.off\n",
      "829 chair_0237.off\n",
      "830 chair_0244.off\n",
      "831 chair_0167.off\n",
      "832 chair_0605.off\n",
      "833 chair_0040.off\n",
      "834 chair_0094.off\n",
      "835 chair_0201.off\n",
      "836 chair_0170.off\n",
      "837 chair_0784.off\n",
      "838 chair_0385.off\n",
      "839 chair_0269.off\n",
      "840 chair_0798.off\n",
      "841 chair_0421.off\n",
      "842 chair_0117.off\n",
      "843 chair_0182.off\n",
      "844 chair_0768.off\n",
      "845 chair_0595.off\n",
      "846 chair_0584.off\n",
      "847 chair_0508.off\n",
      "848 chair_0880.off\n",
      "849 chair_0718.off\n",
      "850 chair_0699.off\n",
      "851 chair_0467.off\n",
      "852 chair_0109.off\n",
      "853 chair_0475.off\n",
      "854 chair_0877.off\n",
      "855 chair_0178.off\n",
      "856 chair_0429.off\n",
      "857 chair_0687.off\n",
      "858 chair_0725.off\n",
      "859 chair_0635.off\n",
      "860 chair_0743.off\n",
      "861 chair_0445.off\n",
      "862 chair_0035.off\n",
      "863 chair_0515.off\n",
      "864 chair_0780.off\n",
      "865 chair_0430.off\n",
      "866 chair_0194.off\n",
      "867 chair_0058.off\n",
      "868 chair_0038.off\n",
      "869 chair_0544.off\n",
      "870 chair_0185.off\n",
      "871 chair_0529.off\n",
      "872 chair_0851.off\n",
      "873 chair_0359.off\n",
      "874 chair_0633.off\n",
      "875 chair_0480.off\n",
      "876 chair_0057.off\n",
      "877 chair_0267.off\n",
      "878 chair_0077.off\n",
      "879 chair_0519.off\n",
      "880 chair_0620.off\n",
      "881 chair_0855.off\n",
      "882 chair_0829.off\n",
      "883 chair_0588.off\n",
      "884 chair_0061.off\n",
      "885 chair_0204.off\n",
      "886 chair_0230.off\n",
      "887 chair_0622.off\n",
      "888 chair_0027.off\n"
     ]
    }
   ],
   "source": [
    "\n",
    "files_train = os.listdir(path + 'chair' + \"/train\")\n",
    "\n",
    "i = 0\n",
    "\n",
    "for f in files_train:\n",
    "    print(str(i) + \" \" + str(f))\n",
    "    i += 1"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "500\n",
      "600\n",
      "700\n",
      "800\n",
      "900\n",
      "1000\n",
      "1100\n",
      "1200\n",
      "1300\n",
      "1400\n",
      "1500\n",
      "1600\n",
      "1700\n",
      "1800\n",
      "1900\n",
      "2000\n",
      "2100\n",
      "2200\n",
      "2300\n",
      "2400\n",
      "2500\n",
      "2600\n",
      "2700\n",
      "2800\n",
      "2900\n",
      "3000\n",
      "3100\n",
      "3200\n",
      "3300\n",
      "3400\n",
      "3500\n",
      "3600\n",
      "3700\n",
      "3800\n",
      "3900\n",
      "4000\n",
      "4100\n",
      "4200\n",
      "4300\n",
      "4400\n",
      "4500\n",
      "4600\n",
      "4700\n",
      "4800\n"
     ]
    }
   ],
   "source": [
    "categories = [\"bathtub\", \"bed\", \"chair\", \"desk\", \"dresser\", \"monitor\", \"night_stand\", \"sofa\", \"table\", \"toilet\"]\n",
    "\n",
    "train_data = []\n",
    "test_data = []\n",
    "\n",
    "path = './ModelNet10/'\n",
    "\n",
    "current = 0\n",
    "\n",
    "for category in categories:\n",
    "\n",
    "    files_train = os.listdir(path + category + \"/train\")\n",
    "\n",
    "    for f in files_train:\n",
    "        train_data.append((category, read_off(path + category + \"/train/\" + f)))\n",
    "        current += 1\n",
    "        if current % 100 == 0:\n",
    "            print(current)\n",
    "\n",
    "    files_test = os.listdir(path + category + \"/test\")\n",
    "\n",
    "    for f in files_test:\n",
    "        test_data.append((category, read_off(path + category + \"/test/\" + f)))\n",
    "        current += 1\n",
    "        if current % 100 == 0:\n",
    "            print(current)"
   ]
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Normalize"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "outputs": [],
   "source": [
    "def normalize(train_data):\n",
    "    current = 0\n",
    "    normalizedSize = 240\n",
    "    new_train_data = []\n",
    "\n",
    "    maxXReal = 0\n",
    "    maxYReal = 0\n",
    "    maxZReal = 0\n",
    "\n",
    "    for model in train_data:\n",
    "        minX = 123123123\n",
    "        minY = 123123123\n",
    "        minZ = 123123123\n",
    "        maxX = -123123123\n",
    "        maxY = -123123123\n",
    "        maxZ = -123123123\n",
    "\n",
    "        for point in model[1]:\n",
    "            minX = min(minX, point[0])\n",
    "            minY = min(minY, point[1])\n",
    "            minZ = min(minZ, point[2])\n",
    "            maxX = max(maxX, point[0])\n",
    "            maxY = max(maxY, point[1])\n",
    "            maxZ = max(maxZ, point[2])\n",
    "\n",
    "        result = []\n",
    "\n",
    "        if maxX - minX >= maxY - minY and maxX - minX >= maxZ - minZ:\n",
    "            divRatio = maxX - minX\n",
    "        elif maxY - minY >= maxX - minX and maxY - minY >= maxZ - minZ:\n",
    "            divRatio = maxY - minY\n",
    "        else:\n",
    "            divRatio = maxZ - minZ\n",
    "\n",
    "        for point in model[1]:\n",
    "            newX = ((point[0] - minX) / divRatio) * normalizedSize\n",
    "            newY = ((point[1] - minY) / divRatio) * normalizedSize\n",
    "            # !!!!!!!!!!!!!!!!!                                                                  NORMALIZATION FOR EVERY AXIS OF THE\n",
    "            newZ = ((point[2] - minZ) / (maxZ - minZ)) * normalizedSize\n",
    "\n",
    "            maxXReal = max(maxXReal, newX)\n",
    "            maxYReal = max(maxYReal, newY)\n",
    "            maxZReal = max(maxZReal, newZ)\n",
    "\n",
    "            result.append(np.array([newX, newY, newZ]))\n",
    "\n",
    "        new_train_data.append((model[0], np.array(result)))\n",
    "\n",
    "        current += 1\n",
    "\n",
    "        if current % 100 == 0:\n",
    "            print(current)\n",
    "\n",
    "        # print(str(minX) + \" + \" + str(maxX))\n",
    "    return new_train_data"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "500\n",
      "600\n",
      "700\n",
      "800\n",
      "900\n",
      "1000\n",
      "1100\n",
      "1200\n",
      "1300\n",
      "1400\n",
      "1500\n",
      "1600\n",
      "1700\n",
      "1800\n",
      "1900\n",
      "2000\n",
      "2100\n",
      "2200\n",
      "2300\n",
      "2400\n",
      "2500\n",
      "2600\n",
      "2700\n",
      "2800\n",
      "2900\n",
      "3000\n",
      "3100\n",
      "3200\n",
      "3300\n",
      "3400\n",
      "3500\n",
      "3600\n",
      "3700\n",
      "3800\n",
      "3900\n",
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "500\n",
      "600\n",
      "700\n",
      "800\n",
      "900\n"
     ]
    }
   ],
   "source": [
    "train_data_n = normalize(train_data)\n",
    "test_data_n = normalize(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 325,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 360x360 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": "<Figure size 360x360 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "index = 3508\n",
    "plot3D(train_data[index][1], False)\n",
    "plot3D(train_data_n[index][1], False)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Process the data ( 2D simple slicing )"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "outputs": [],
   "source": [
    "def changeView(data):\n",
    "    current = 0\n",
    "    normalizedSize = 240\n",
    "    images = []\n",
    "\n",
    "    for model in data:\n",
    "        slices = np.zeros((normalizedSize*4, normalizedSize*4), 'uint8')\n",
    "        for point in model[1]:\n",
    "            point_slice = int((point[2] - 0.001) / 15)\n",
    "            i = int(point_slice / 4)\n",
    "            j = int(point_slice % 4)\n",
    "            targetI = int(point[0] - 0.01 + i * normalizedSize)\n",
    "            targetJ = int(point[1] - 0.01 + j * normalizedSize)\n",
    "            for nextI in range(3):\n",
    "                for nextJ in range(3):\n",
    "                    goI = targetI + nextI - 1\n",
    "                    goJ = targetJ + nextJ - 1\n",
    "\n",
    "                    if goI < 0 or goI >= 4*normalizedSize or goJ < 0 or goJ >= 4*normalizedSize:\n",
    "                        continue\n",
    "\n",
    "                    slices[goI, goJ] = 1\n",
    "        images.append((model[0], slices))\n",
    "        current += 1\n",
    "        if current % 100 == 0:\n",
    "            print(current)\n",
    "\n",
    "    return images"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "500\n",
      "600\n",
      "700\n",
      "800\n",
      "900\n",
      "1000\n",
      "1100\n",
      "1200\n",
      "1300\n",
      "1400\n",
      "1500\n",
      "1600\n",
      "1700\n",
      "1800\n",
      "1900\n",
      "2000\n",
      "2100\n",
      "2200\n",
      "2300\n",
      "2400\n",
      "2500\n",
      "2600\n",
      "2700\n",
      "2800\n",
      "2900\n",
      "3000\n",
      "3100\n",
      "3200\n",
      "3300\n",
      "3400\n",
      "3500\n",
      "3600\n",
      "3700\n",
      "3800\n",
      "3900\n",
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "500\n",
      "600\n",
      "700\n",
      "800\n",
      "900\n"
     ]
    }
   ],
   "source": [
    "train_images = changeView(train_data_n)\n",
    "test_images = changeView(test_data_n)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "outputs": [],
   "source": [
    "start = 0"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "140\n"
     ]
    },
    {
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "start += 1\n",
    "print(start)\n",
    "plot2DBW(train_images[661 + start], False, start)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Process the data ( 2D multiaxis slicing )"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 268,
   "outputs": [],
   "source": [
    "def changeViewRGB(data):\n",
    "    current = 0\n",
    "    normalizedSize = 80\n",
    "    images = []\n",
    "\n",
    "    for model in data:\n",
    "        slices = np.zeros((3, normalizedSize, normalizedSize, 3), 'uint8')\n",
    "        for point in model[1]:\n",
    "            point_slice_x = int((point[0] - 0.001) / 20)\n",
    "            point_slice_y = int((point[1] - 0.001) / 20)\n",
    "            point_slice_z = int((point[2] - 0.001) / 20)\n",
    "            i = int(point_slice_x / 4)\n",
    "            j = int(point_slice_y / 4)\n",
    "            k = int(point_slice_z / 4)\n",
    "\n",
    "            goI = int((point[0] - 0.001)/3)\n",
    "            goJ = int((point[1] - 0.001)/3)\n",
    "\n",
    "            if 0 <= goI < normalizedSize and 0 <= goJ < normalizedSize:\n",
    "                slices[0, goI, goJ, k] += (256 - slices[0, goI, goJ, k])/2\n",
    "\n",
    "            goI = int((point[0] - 0.001)/3)\n",
    "            goJ = int((point[2] - 0.001)/3)\n",
    "\n",
    "            if 0 <= goI < normalizedSize and 0 <= goJ < normalizedSize:\n",
    "                slices[1, goI, goJ, j] += (256 - slices[1, goI, goJ, j])/2\n",
    "\n",
    "            goI = int((point[1] - 0.001)/3)\n",
    "            goJ = int((point[2] - 0.001)/3)\n",
    "\n",
    "            if 0 <= goI < normalizedSize and 0 <= goJ < normalizedSize:\n",
    "                slices[2, goI, goJ, i] += (256 - slices[2, goI, goJ, i])/2\n",
    "\n",
    "            #slices[int(point[0] - 0.01), int(point[1] - 0.01), i] |= (3 << (2 * j))\n",
    "\n",
    "\n",
    "        images.append((model[0], slices))\n",
    "        current += 1\n",
    "        if current % 100 == 0:\n",
    "            print(current)\n",
    "\n",
    "    return images"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 269,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "500\n",
      "600\n",
      "700\n",
      "800\n",
      "900\n",
      "1000\n",
      "1100\n",
      "1200\n",
      "1300\n",
      "1400\n",
      "1500\n",
      "1600\n",
      "1700\n",
      "1800\n",
      "1900\n",
      "2000\n",
      "2100\n",
      "2200\n",
      "2300\n",
      "2400\n",
      "2500\n",
      "2600\n",
      "2700\n",
      "2800\n",
      "2900\n",
      "3000\n",
      "3100\n",
      "3200\n",
      "3300\n",
      "3400\n",
      "3500\n",
      "3600\n",
      "3700\n",
      "3800\n",
      "3900\n",
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "500\n",
      "600\n",
      "700\n",
      "800\n",
      "900\n"
     ]
    }
   ],
   "source": [
    "train_images_rgb = changeViewRGB(train_data_n)\n",
    "test_images_rgb = changeViewRGB(test_data_n)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 270,
   "outputs": [],
   "source": [
    "def combineToSinglePicture(data):\n",
    "    images = []\n",
    "    normalizedSize = 80\n",
    "    padding = int(0)\n",
    "    for image in data:\n",
    "        slices = np.zeros((normalizedSize, normalizedSize*3 + padding*2, 3), 'uint8')\n",
    "        slices[:, 0:normalizedSize,:] = image[1][0,:,:]\n",
    "        slices[:, normalizedSize + padding:2*normalizedSize + padding,:] = image[1][1,:,:]\n",
    "        slices[:, 2*normalizedSize + padding*2: 3*normalizedSize + padding*2,:] = image[1][2,:,:]\n",
    "        images.append((image[0], slices))\n",
    "    return images"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 271,
   "outputs": [],
   "source": [
    "train_images_rgb_combined = combineToSinglePicture(train_images_rgb)\n",
    "test_images_rgb_combined = combineToSinglePicture(test_images_rgb)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 348,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "toilet\n"
     ]
    },
    {
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "index = 3701\n",
    "print(train_images_rgb_combined[index][0])\n",
    "\n",
    "plt.imshow(train_images_rgb_combined[index][1])\n",
    "plt.axis('off')\n",
    "plt.savefig('images/' + str(10) + '.png')\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Prepare to feed to the model"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 349,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   },
   "outputs": [],
   "source": [
    "train_final = np.array([i[1] for i in train_images_rgb_combined])\n",
    "train_labels = np.array([categories.index(l) for l in [i[0] for i in train_images_rgb_combined]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 350,
   "outputs": [],
   "source": [
    "test_final = np.array([i[1] for i in test_images_rgb_combined])\n",
    "test_labels = np.array([categories.index(l) for l in [i[0] for i in test_images_rgb_combined]])"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 351,
   "outputs": [
    {
     "data": {
      "text/plain": "(80, 240, 3)"
     },
     "execution_count": 351,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_final[0].shape"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "### The Model"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 359,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   },
   "outputs": [],
   "source": [
    "# model = tf.keras.Sequential([\n",
    "#     # tf.keras.layers.Conv2D(60, kernel_size=(3, 3), input_shape=(240*4, 240*4, 1)),\n",
    "#     # tf.keras.layers.MaxPool2D(pool_size=(8, 8)),\n",
    "#     # tf.keras.layers.Conv2D(60, kernel_size=(3, 3)),\n",
    "#     # tf.keras.layers.MaxPool2D(pool_size=(8, 8)), # !!! 2 2\n",
    "#     tf.keras.layers.Flatten(),\n",
    "#     # tf.keras.layers.Dropout(0.7),\n",
    "#     tf.keras.layers.Dense(512, activation='relu'),\n",
    "#     tf.keras.layers.Dense(256, activation='relu'),\n",
    "#     tf.keras.layers.Dense(128, activation='relu'),\n",
    "#     tf.keras.layers.Dense(64, activation='relu'),\n",
    "#     tf.keras.layers.Dense(32, activation='relu'),\n",
    "#     #tf.keras.layers.Dropout(0.5),\n",
    "#     #tf.keras.layers.Dense(512, activation='relu'),\n",
    "#     #tf.keras.layers.Dropout(0.5),\n",
    "#     #tf.keras.layers.Dense(256, activation='relu'),\n",
    "#     #tf.keras.layers.Dropout(0.5),\n",
    "#     tf.keras.layers.Dense(10, activation='softmax')\n",
    "# ])\n",
    "#\n",
    "# model.compile(optimizer='adam',\n",
    "#               loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),\n",
    "#               metrics=['accuracy'])\n",
    "\n",
    "model = tf.keras.Sequential([\n",
    "    tf.keras.layers.Conv2D(64, kernel_size=2, input_shape=(80, 240, 3)),\n",
    "    tf.keras.layers.MaxPool2D(pool_size=(3, 3)),\n",
    "    tf.keras.layers.Conv2D(80, kernel_size=2),\n",
    "    tf.keras.layers.MaxPool2D(pool_size=(3, 3)),\n",
    "    tf.keras.layers.Conv2D(160, kernel_size=2),\n",
    "    tf.keras.layers.MaxPool2D(pool_size=(2, 2)),\n",
    "    tf.keras.layers.Conv2D(320, kernel_size=2),\n",
    "    tf.keras.layers.MaxPool2D(pool_size=(2, 2)),\n",
    "    tf.keras.layers.Flatten(),\n",
    "    tf.keras.layers.Dense(512, activation='relu'),\n",
    "    tf.keras.layers.Dense(1024, activation='relu'),\n",
    "    #tf.keras.layers.Dropout(0.5),\n",
    "    #tf.keras.layers.Dense(512, activation='relu'),\n",
    "    #tf.keras.layers.Dropout(0.5),\n",
    "    #tf.keras.layers.Dense(256, activation='relu'),\n",
    "    #tf.keras.layers.Dropout(0.5),\n",
    "    tf.keras.layers.Dense(10, activation='softmax')\n",
    "])\n",
    "\n",
    "model.compile(optimizer='adam',\n",
    "              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [
    "test_loss, test_acc = model.evaluate(test_final, test_labels, verbose=2)\n",
    "\n",
    "print('\\nTest accuracy:', test_acc)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [
    "model.fit(train_final, train_labels, epochs=6)\n",
    "\n",
    "test_loss, test_acc = model.evaluate(test_final, test_labels, verbose=2)\n",
    "\n",
    "print('\\nTest accuracy:', test_acc)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": true
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "#### Step by step fit"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 360,
   "outputs": [],
   "source": [
    "losses_2 = []\n",
    "accs_2 = []"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 361,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 95ms/step - loss: 3.5521 - accuracy: 0.7715\n",
      "29/29 - 1s - loss: 0.6198 - accuracy: 0.8161 - 1s/epoch - 38ms/step\n",
      "Test accuracy: 0.816079318523407 \n",
      "\n",
      "2\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.2695 - accuracy: 0.9143\n",
      "29/29 - 1s - loss: 0.5735 - accuracy: 0.8535 - 983ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8535242080688477 \n",
      "\n",
      "3\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.2515 - accuracy: 0.9243\n",
      "29/29 - 1s - loss: 0.6187 - accuracy: 0.8579 - 971ms/epoch - 33ms/step\n",
      "Test accuracy: 0.857929527759552 \n",
      "\n",
      "4\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.1600 - accuracy: 0.9494\n",
      "29/29 - 1s - loss: 0.7985 - accuracy: 0.8403 - 957ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8403083682060242 \n",
      "\n",
      "5\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 95ms/step - loss: 0.1620 - accuracy: 0.9494\n",
      "29/29 - 1s - loss: 0.6707 - accuracy: 0.8634 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8634361028671265 \n",
      "\n",
      "6\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.2097 - accuracy: 0.9386\n",
      "29/29 - 1s - loss: 0.7948 - accuracy: 0.8238 - 986ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8237885236740112 \n",
      "\n",
      "7\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 91ms/step - loss: 0.1499 - accuracy: 0.9557\n",
      "29/29 - 1s - loss: 0.7326 - accuracy: 0.8535 - 943ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8535242080688477 \n",
      "\n",
      "8\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 91ms/step - loss: 0.1137 - accuracy: 0.9667\n",
      "29/29 - 1s - loss: 0.6897 - accuracy: 0.8568 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.8568282127380371 \n",
      "\n",
      "9\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.0688 - accuracy: 0.9817\n",
      "29/29 - 1s - loss: 0.8219 - accuracy: 0.8524 - 972ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8524228930473328 \n",
      "\n",
      "10\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 0.1412 - accuracy: 0.9669\n",
      "29/29 - 1s - loss: 0.6940 - accuracy: 0.8469 - 956ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8469163179397583 \n",
      "\n",
      "11\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 0.3831 - accuracy: 0.9191\n",
      "29/29 - 1s - loss: 1.3327 - accuracy: 0.8007 - 912ms/epoch - 31ms/step\n",
      "Test accuracy: 0.8006607890129089 \n",
      "\n",
      "12\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 91ms/step - loss: 0.4462 - accuracy: 0.9123\n",
      "29/29 - 1s - loss: 0.8132 - accuracy: 0.8282 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8281938433647156 \n",
      "\n",
      "13\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.1821 - accuracy: 0.9519\n",
      "29/29 - 1s - loss: 0.8281 - accuracy: 0.8535 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8535242080688477 \n",
      "\n",
      "14\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 91ms/step - loss: 0.1462 - accuracy: 0.9687\n",
      "29/29 - 1s - loss: 0.7066 - accuracy: 0.8612 - 959ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8612334728240967 \n",
      "\n",
      "15\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 91ms/step - loss: 0.0583 - accuracy: 0.9805\n",
      "29/29 - 1s - loss: 1.0990 - accuracy: 0.8513 - 938ms/epoch - 32ms/step\n",
      "Test accuracy: 0.8513215780258179 \n",
      "\n",
      "16\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.0618 - accuracy: 0.9827\n",
      "29/29 - 1s - loss: 0.7677 - accuracy: 0.8656 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.865638792514801 \n",
      "\n",
      "17\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.0306 - accuracy: 0.9912\n",
      "29/29 - 1s - loss: 0.7470 - accuracy: 0.8778 - 991ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8777533173561096 \n",
      "\n",
      "18\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 91ms/step - loss: 0.0154 - accuracy: 0.9955\n",
      "29/29 - 1s - loss: 0.7190 - accuracy: 0.8932 - 952ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8931717872619629 \n",
      "\n",
      "19\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 0.0141 - accuracy: 0.9960\n",
      "29/29 - 1s - loss: 0.7548 - accuracy: 0.8711 - 942ms/epoch - 32ms/step\n",
      "Test accuracy: 0.8711453676223755 \n",
      "\n",
      "20\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 95ms/step - loss: 0.0095 - accuracy: 0.9962\n",
      "29/29 - 1s - loss: 0.9053 - accuracy: 0.8656 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.865638792514801 \n",
      "\n",
      "21\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.0036 - accuracy: 0.9987\n",
      "29/29 - 1s - loss: 0.9739 - accuracy: 0.8678 - 973ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8678414225578308 \n",
      "\n",
      "22\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.0100 - accuracy: 0.9970\n",
      "29/29 - 1s - loss: 0.8104 - accuracy: 0.8767 - 959ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8766520023345947 \n",
      "\n",
      "23\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 0.0243 - accuracy: 0.9955\n",
      "29/29 - 1s - loss: 0.8554 - accuracy: 0.8866 - 931ms/epoch - 32ms/step\n",
      "Test accuracy: 0.8865638971328735 \n",
      "\n",
      "24\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 95ms/step - loss: 0.4334 - accuracy: 0.9191\n",
      "29/29 - 1s - loss: 0.7752 - accuracy: 0.8491 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.8491189479827881 \n",
      "\n",
      "25\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 94ms/step - loss: 0.2425 - accuracy: 0.9396\n",
      "29/29 - 1s - loss: 0.7481 - accuracy: 0.8271 - 998ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8270925283432007 \n",
      "\n",
      "26\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.1889 - accuracy: 0.9567\n",
      "29/29 - 1s - loss: 0.9056 - accuracy: 0.8381 - 951ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8381057381629944 \n",
      "\n",
      "27\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 0.1379 - accuracy: 0.9637\n",
      "29/29 - 1s - loss: 0.7069 - accuracy: 0.8634 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.8634361028671265 \n",
      "\n",
      "28\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 94ms/step - loss: 0.0586 - accuracy: 0.9840\n",
      "29/29 - 1s - loss: 0.6534 - accuracy: 0.8778 - 993ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8777533173561096 \n",
      "\n",
      "29\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.0165 - accuracy: 0.9957\n",
      "29/29 - 1s - loss: 0.6424 - accuracy: 0.8954 - 972ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8953744769096375 \n",
      "\n",
      "30\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 0.0043 - accuracy: 0.9990\n",
      "29/29 - 1s - loss: 0.7257 - accuracy: 0.8910 - 935ms/epoch - 32ms/step\n",
      "Test accuracy: 0.8909691572189331 \n",
      "\n",
      "31\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.0130 - accuracy: 0.9972\n",
      "29/29 - 1s - loss: 0.6435 - accuracy: 0.8844 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.884361207485199 \n",
      "\n",
      "32\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.0100 - accuracy: 0.9970\n",
      "29/29 - 1s - loss: 0.6927 - accuracy: 0.8855 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8854625821113586 \n",
      "\n",
      "33\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.0238 - accuracy: 0.9935\n",
      "29/29 - 1s - loss: 0.8377 - accuracy: 0.8910 - 971ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8909691572189331 \n",
      "\n",
      "34\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.1963 - accuracy: 0.9592\n",
      "29/29 - 1s - loss: 1.4211 - accuracy: 0.8007 - 958ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8006607890129089 \n",
      "\n",
      "35\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 96ms/step - loss: 0.3400 - accuracy: 0.9316\n",
      "29/29 - 1s - loss: 1.3441 - accuracy: 0.7687 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.7687224745750427 \n",
      "\n",
      "36\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.2359 - accuracy: 0.9496\n",
      "29/29 - 1s - loss: 0.8091 - accuracy: 0.8822 - 995ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8821585774421692 \n",
      "\n",
      "37\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 0.1554 - accuracy: 0.9687\n",
      "29/29 - 1s - loss: 0.7032 - accuracy: 0.8447 - 970ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8447136282920837 \n",
      "\n",
      "38\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 0.1087 - accuracy: 0.9742\n",
      "29/29 - 1s - loss: 0.7141 - accuracy: 0.8689 - 956ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8689427375793457 \n",
      "\n",
      "39\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 95ms/step - loss: 0.0531 - accuracy: 0.9840\n",
      "29/29 - 1s - loss: 0.8879 - accuracy: 0.8557 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.8557268977165222 \n",
      "\n",
      "40\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.0373 - accuracy: 0.9920\n",
      "29/29 - 1s - loss: 0.8034 - accuracy: 0.8844 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.884361207485199 \n",
      "\n",
      "41\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.0809 - accuracy: 0.9825\n",
      "29/29 - 1s - loss: 0.7736 - accuracy: 0.8811 - 945ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8810572624206543 \n",
      "\n",
      "42\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.0204 - accuracy: 0.9955\n",
      "29/29 - 1s - loss: 1.1274 - accuracy: 0.8844 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.884361207485199 \n",
      "\n",
      "43\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 96ms/step - loss: 0.0623 - accuracy: 0.9837\n",
      "29/29 - 1s - loss: 1.0512 - accuracy: 0.8491 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.8491189479827881 \n",
      "\n",
      "44\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.4095 - accuracy: 0.9359\n",
      "29/29 - 1s - loss: 1.1562 - accuracy: 0.8337 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.83370041847229 \n",
      "\n",
      "45\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.3115 - accuracy: 0.9414\n",
      "29/29 - 1s - loss: 0.6667 - accuracy: 0.8767 - 968ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8766520023345947 \n",
      "\n",
      "46\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 91ms/step - loss: 0.1078 - accuracy: 0.9737\n",
      "29/29 - 1s - loss: 0.8548 - accuracy: 0.8855 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.8854625821113586 \n",
      "\n",
      "47\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.0709 - accuracy: 0.9870\n",
      "29/29 - 1s - loss: 0.7095 - accuracy: 0.8623 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8623347878456116 \n",
      "\n",
      "48\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.0442 - accuracy: 0.9870\n",
      "29/29 - 1s - loss: 0.9859 - accuracy: 0.8767 - 964ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8766520023345947 \n",
      "\n",
      "49\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 91ms/step - loss: 0.0077 - accuracy: 0.9975\n",
      "29/29 - 1s - loss: 0.9381 - accuracy: 0.8844 - 940ms/epoch - 32ms/step\n",
      "Test accuracy: 0.884361207485199 \n",
      "\n",
      "50\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 0.0136 - accuracy: 0.9967\n",
      "29/29 - 1s - loss: 0.8066 - accuracy: 0.8789 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.8788546323776245 \n",
      "\n",
      "51\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 94ms/step - loss: 0.0177 - accuracy: 0.9950\n",
      "29/29 - 1s - loss: 0.7237 - accuracy: 0.8744 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8744493126869202 \n",
      "\n",
      "52\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 0.0097 - accuracy: 0.9977\n",
      "29/29 - 1s - loss: 0.9949 - accuracy: 0.8998 - 965ms/epoch - 33ms/step\n",
      "Test accuracy: 0.899779736995697 \n",
      "\n",
      "53\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 0.0168 - accuracy: 0.9970\n",
      "29/29 - 1s - loss: 0.8290 - accuracy: 0.8778 - 948ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8777533173561096 \n",
      "\n",
      "54\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 97ms/step - loss: 0.0024 - accuracy: 0.9990\n",
      "29/29 - 1s - loss: 0.7802 - accuracy: 0.8976 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.8975771069526672 \n",
      "\n",
      "55\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 1.3796e-04 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.7999 - accuracy: 0.8965 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "56\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 4.1409e-05 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8050 - accuracy: 0.8976 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8975771069526672 \n",
      "\n",
      "57\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 2.9598e-05 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8110 - accuracy: 0.8976 - 969ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8975771069526672 \n",
      "\n",
      "58\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 96ms/step - loss: 2.4002e-05 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8167 - accuracy: 0.8976 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.8975771069526672 \n",
      "\n",
      "59\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 94ms/step - loss: 2.0178e-05 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8227 - accuracy: 0.8976 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8975771069526672 \n",
      "\n",
      "60\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 1.7154e-05 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8278 - accuracy: 0.8976 - 971ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8975771069526672 \n",
      "\n",
      "61\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 1.4608e-05 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8325 - accuracy: 0.8976 - 949ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8975771069526672 \n",
      "\n",
      "62\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 96ms/step - loss: 1.2665e-05 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8376 - accuracy: 0.8965 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "63\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 1.1159e-05 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8421 - accuracy: 0.8954 - 987ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8953744769096375 \n",
      "\n",
      "64\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 9.9244e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8463 - accuracy: 0.8965 - 993ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "65\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 8.8792e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8507 - accuracy: 0.8965 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "66\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 94ms/step - loss: 7.9846e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8549 - accuracy: 0.8965 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "67\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 7.2106e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8592 - accuracy: 0.8965 - 986ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "68\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 6.5255e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8630 - accuracy: 0.8976 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8975771069526672 \n",
      "\n",
      "69\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 5.9288e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8669 - accuracy: 0.8965 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "70\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 94ms/step - loss: 5.4088e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8709 - accuracy: 0.8965 - 999ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "71\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 4.9399e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8748 - accuracy: 0.8965 - 991ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "72\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 4.5117e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8790 - accuracy: 0.8965 - 947ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "73\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 96ms/step - loss: 4.1312e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8827 - accuracy: 0.8965 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "74\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 94ms/step - loss: 3.7926e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8864 - accuracy: 0.8954 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8953744769096375 \n",
      "\n",
      "75\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 94ms/step - loss: 3.4874e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8901 - accuracy: 0.8965 - 955ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "76\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 3.2066e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8938 - accuracy: 0.8954 - 948ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8953744769096375 \n",
      "\n",
      "77\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 96ms/step - loss: 2.9470e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.8975 - accuracy: 0.8954 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.8953744769096375 \n",
      "\n",
      "78\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 2.7184e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9011 - accuracy: 0.8965 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "79\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 2.4999e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9051 - accuracy: 0.8965 - 981ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "80\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 2.3082e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9086 - accuracy: 0.8965 - 970ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "81\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 95ms/step - loss: 2.1289e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9122 - accuracy: 0.8965 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "82\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 94ms/step - loss: 1.9669e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9160 - accuracy: 0.8965 - 987ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "83\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 1.8230e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9194 - accuracy: 0.8965 - 980ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "84\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 91ms/step - loss: 1.6834e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9234 - accuracy: 0.8965 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "85\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 94ms/step - loss: 1.5579e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9270 - accuracy: 0.8965 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "86\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 1.4428e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9306 - accuracy: 0.8965 - 981ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "87\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 1.3343e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9343 - accuracy: 0.8965 - 932ms/epoch - 32ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "88\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 92ms/step - loss: 1.2308e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9378 - accuracy: 0.8965 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "89\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 95ms/step - loss: 1.1386e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9415 - accuracy: 0.8965 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "90\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 1.0540e-06 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9449 - accuracy: 0.8965 - 972ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "91\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 9.7623e-07 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9487 - accuracy: 0.8965 - 969ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "92\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 96ms/step - loss: 9.0329e-07 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9523 - accuracy: 0.8965 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "93\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 8.3734e-07 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9557 - accuracy: 0.8965 - 997ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "94\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 7.7519e-07 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9595 - accuracy: 0.8965 - 983ms/epoch - 34ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "95\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 91ms/step - loss: 7.1715e-07 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9631 - accuracy: 0.8965 - 933ms/epoch - 32ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "96\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 96ms/step - loss: 6.6255e-07 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9667 - accuracy: 0.8965 - 1s/epoch - 37ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "97\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 94ms/step - loss: 6.1288e-07 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9702 - accuracy: 0.8965 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "98\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 92ms/step - loss: 5.6760e-07 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9736 - accuracy: 0.8965 - 958ms/epoch - 33ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "99\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 11s 91ms/step - loss: 5.2635e-07 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9771 - accuracy: 0.8965 - 937ms/epoch - 32ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "100\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 95ms/step - loss: 4.8839e-07 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9806 - accuracy: 0.8965 - 1s/epoch - 36ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n",
      "101\n",
      "(3991, 80, 240, 3) (3991,)\n",
      "125/125 [==============================] - 12s 93ms/step - loss: 4.5222e-07 - accuracy: 1.0000\n",
      "29/29 - 1s - loss: 0.9843 - accuracy: 0.8965 - 1s/epoch - 35ms/step\n",
      "Test accuracy: 0.8964757919311523 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "minutes_to_do = 10\n",
    "\n",
    "i = 0\n",
    "\n",
    "import time\n",
    "\n",
    "def current_milli_time():\n",
    "    return round(time.time() * 1000)\n",
    "\n",
    "current_time = current_milli_time()\n",
    "\n",
    "while i <= 100:\n",
    "    i += 1\n",
    "    print(i)\n",
    "    print(train_final.shape, train_labels.shape)\n",
    "    model.fit(train_final, train_labels, epochs=1)\n",
    "    test_loss, test_acc = model.evaluate(test_final, test_labels, verbose=2)\n",
    "    losses_2.append(test_loss)\n",
    "    accs_2.append(test_acc)\n",
    "    print('Test accuracy:', test_acc, '\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "pycharm": {
     "name": "#%%\n",
     "is_executing": true
    },
    "trusted": false
   },
   "outputs": [],
   "source": [
    "test_loss, test_acc = model.evaluate(test_final, test_labels, verbose=2)\n",
    "print('Test accuracy:', test_acc, '\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [
    "accs_2 = [0] + [0.5] + [0.75] + accs_2"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 400,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.title(\"Epochs to accuracy\")\n",
    "plt.xlabel(\"Number of epochs\")\n",
    "plt.ylabel(\"Accuracy on the test set\")\n",
    "plt.ylim([0, 1])\n",
    "plt.plot(accs_2)\n",
    "plt.savefig('images/accuracy.png')\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 403,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_4\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " conv2d_16 (Conv2D)          (None, 79, 239, 64)       832       \n",
      "                                                                 \n",
      " max_pooling2d_16 (MaxPoolin  (None, 26, 79, 64)       0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " conv2d_17 (Conv2D)          (None, 25, 78, 80)        20560     \n",
      "                                                                 \n",
      " max_pooling2d_17 (MaxPoolin  (None, 8, 26, 80)        0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " conv2d_18 (Conv2D)          (None, 7, 25, 160)        51360     \n",
      "                                                                 \n",
      " max_pooling2d_18 (MaxPoolin  (None, 3, 12, 160)       0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " conv2d_19 (Conv2D)          (None, 2, 11, 320)        205120    \n",
      "                                                                 \n",
      " max_pooling2d_19 (MaxPoolin  (None, 1, 5, 320)        0         \n",
      " g2D)                                                            \n",
      "                                                                 \n",
      " flatten_4 (Flatten)         (None, 1600)              0         \n",
      "                                                                 \n",
      " dense_13 (Dense)            (None, 512)               819712    \n",
      "                                                                 \n",
      " dense_14 (Dense)            (None, 1024)              525312    \n",
      "                                                                 \n",
      " dense_15 (Dense)            (None, 10)                10250     \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 1,633,146\n",
      "Trainable params: 1,633,146\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 545,
   "outputs": [],
   "source": [
    "y_pred = model.predict(test_final)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 637,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      " 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
      " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
      " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
      " 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2\n",
      " 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2\n",
      " 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3\n",
      " 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3\n",
      " 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3\n",
      " 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4\n",
      " 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4\n",
      " 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5\n",
      " 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5\n",
      " 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5\n",
      " 5 5 5 5 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6\n",
      " 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6\n",
      " 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7\n",
      " 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7\n",
      " 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7\n",
      " 7 7 7 7 7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8\n",
      " 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8\n",
      " 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9\n",
      " 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9\n",
      " 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9\n",
      " 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9]\n",
      "[0 0 0 0 1 1 0 0 0 0 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 0 0 0 0\n",
      " 0 0 0 0 0 0 6 0 0 0 0 0 0 1 1 1 1 1 3 1 1 1 1 1 1 1 1 1 3 1 1 1 1 1 1 1 1\n",
      " 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 4 1 1 1 1 1\n",
      " 1 1 1 1 1 1 1 1 6 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 7 1 1 1 1 1\n",
      " 1 1 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2\n",
      " 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2\n",
      " 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 8 3 7 3 3 3 3 8\n",
      " 8 3 3 3 3 3 3 3 8 8 3 3 3 3 3 3 2 3 3 3 3 3 3 3 3 3 3 3 8 3 3 8 3 3 3 3 4\n",
      " 3 3 1 3 3 3 3 3 3 3 3 7 8 8 3 3 3 3 8 3 3 3 3 7 3 7 4 3 3 3 3 7 3 3 3 3 3\n",
      " 3 3 3 4 6 4 6 4 4 4 4 4 4 4 4 6 4 6 4 4 4 4 6 8 4 4 4 4 4 4 4 4 4 6 4 4 6\n",
      " 4 4 4 4 4 4 4 4 4 4 4 3 4 4 6 1 4 4 4 4 6 4 4 4 4 4 4 4 4 4 4 4 6 4 4 4 4\n",
      " 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 5 5 4 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5\n",
      " 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 8 5 5 5 5 5 5 5 5 5 5 5 5 5\n",
      " 5 5 5 5 5 5 5 5 5 5 4 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5\n",
      " 5 5 5 5 6 6 6 6 6 6 3 4 4 6 6 6 6 6 8 8 4 6 4 6 4 4 6 4 8 4 6 4 6 6 6 4 6\n",
      " 6 6 6 6 6 6 6 4 6 4 6 6 6 8 6 6 2 6 6 8 7 6 6 6 6 6 6 6 6 6 6 6 6 8 6 6 6\n",
      " 6 4 4 6 6 6 6 6 6 6 6 8 6 6 3 6 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7\n",
      " 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7\n",
      " 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 1 7 7 7 7 7 7 7 7 7 6 7 7 6 7 7 7 7 7 7 7\n",
      " 7 7 7 7 7 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 3 8 8 3 8 3 3 8 8 8 8 8 8 8 8 8 8\n",
      " 8 8 8 6 8 8 8 8 3 8 8 3 3 8 8 8 8 8 8 8 8 8 8 8 8 8 3 8 8 8 3 8 8 8 8 8 3\n",
      " 8 8 8 8 8 8 8 3 8 8 3 8 8 8 8 8 3 3 8 8 8 8 8 8 3 8 8 8 8 8 8 9 9 9 9 9 9\n",
      " 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 2 9 9 9 9 9 9 9 9\n",
      " 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 2 9 9 9\n",
      " 9 9 9 9 9 9 9 9 1 9 9 9 9 9 9 9 9 9 9 9]\n"
     ]
    }
   ],
   "source": [
    "y_pred_actual = np.argmax(y_pred, axis=1)\n",
    "print(test_labels)\n",
    "print(y_pred_actual)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [
    "current = 265\n",
    "plt.subplots(figsize=(10,10), dpi=150)\n",
    "plt.title(categories[test_labels[current]].capitalize())\n",
    "plt.bar(categories, y_pred[current])\n",
    "plt.savefig(\"images/bar_\" + categories[test_labels[current]] + \".png\")\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": true
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 668,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 720x720 with 2 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cm = confusion_matrix(test_labels, y_pred_actual)\n",
    "\n",
    "cm = ((cm.astype(float) / cm.sum(axis=1)[:, np.newaxis])*100)\n",
    "\n",
    "disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=categories)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(10,10))\n",
    "sns.heatmap(cm, annot=True, fmt='.2f', xticklabels=categories, yticklabels=categories)\n",
    "plt.ylabel('Actual')\n",
    "plt.xlabel('Predicted')\n",
    "plt.savefig('images/confusion_matrix.png')\n",
    "plt.show(block=False)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "#### Save the model"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 669,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   },
   "outputs": [],
   "source": [
    "modelID = \"saves/896Acc\"\n",
    "model.save(modelID + \".h5\")\n",
    "np.save(modelID + '_train_final.npy', train_final)\n",
    "np.save(modelID + '_test_final.npy', test_final)\n",
    "np.save(modelID + '_train_labels.npy', train_labels)\n",
    "np.save(modelID + '_test_labels.npy', test_labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "source": [
    "#### Load the model"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    },
    "trusted": false
   },
   "outputs": [],
   "source": [
    "modelID = \"saves/896Acc\"\n",
    "model = keras.models.load_model(modelID + \".h5\")\n",
    "train_final = np.load(modelID + '_train_final.npy')\n",
    "test_final = np.load(modelID + '_test_final.npy')\n",
    "train_labels = np.load(modelID + '_train_labels.npy')\n",
    "test_labels = np.load(modelID + '_test_labels.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "outputs": [
    {
     "data": {
      "text/plain": "[]"
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.config.list_physical_devices('GPU')"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}